For me, 2025 was the year of AI advancing science. I came to this through weather forecasting, helping to launch WeatherNext in Google Cloud. As the year progressed, we learned more about AI’s potential to accelerate nearly every scientific discipline.

Modern science relies on large datasets and scaled data analysis. General purpose models like Gemini handle both structured and unstructured scientific sources, making them perfect for agentic AI systems like AI Co-Scientist.

In Co-Scientist, agents search and understand scientific literature and datasets. Other agents generate novel research ideas based on this research combined with the scientist’s prompt. Still other agents review generated ideas and score them for novelty, feasibility, and impact. Similar to LMArena, Co-Scientist runs a tournament with AI agent judges, simulating peer review. This produces ELO scores that let the most promising research ideas bubble to the top.

This process accomplishes research that would otherwise take scientists days or weeks. One colleague told me, referring to AI agents for science: “this really works.” If you know any scientists, that’s high praise.

In addition to language, video, code, and other modalities, AI is now transforming science in significant ways. This includes advancements in protein folding, materials science, and weather forecasting that are accelerating as we head into 2025.

In the AI-driven weather space, Google DeepMind’s WeatherNext Graph and Gen AI models predict global weather variables with state-of-the-art accuracy up to 15 days in advance. Unlike traditional numerical weather prediction systems, WeatherNext models learn from historical weather observations rather than predicting the weather with hand-crafted physics formulas. The probabilistic WeatherNext Gen model also uses diffusion, similar to the architectures used by Imagen and Veo. WeatherNext models are faster and more efficient than traditional physics-based weather models and yield superior forecast reliability.

What does this mean for you and your business? Well, almost every industry and community is affected by the weather. If you’re working in industries like agriculture, manufacturing, logistics, energy, and financial services, it’s worth checking out what these new AI weather forecasting models can do to improve your day to day operations. Imagine knowing about severe weather events days earlier and with better accuracy – powerful stuff.

What specific skills and knowledge gaps need to be addressed to work effectively with AI?

Jeff Sternberg, Technical Director, Office of the CTO

In addition to the skills gaps addressed by my colleagues here, rolling out a successful learning program for AI is also critical. It’s similar to any enterprise learning and development program. However, it differs in a few key ways:

• Velocity of change: AI is changing extremely fast, so learning materials will become quickly out of date. What’s the fix? Plan to evolve the learning program quickly – rather than doing a yearly refresh of materials, do this semi-annually or even quarterly.
• Inclusivity: While AI is intrinsically technical, the best outcomes will happen if the learning program is extended to every team and role in the organization, including the business functions. In OCTO, we have worked with CTOs and CEOs that are driving this change through every business unit and seeing great results in getting the entire organization speaking the same AI language.
• Multi-modality: Technical teams are used to engaging with new technology using APIs, programming languages, and similar. Gen AI communicates through text, video, audio, and more. For software engineers, that means using skills like communication through natural language (e.g. prompts), which is both new and exciting!

What are the biggest real-world obstacles companies face when trying to implement and scale AI solutions?

Jeff Sternberg, Technical Director, Office of the CTO

Generative AI can increase implementation complexity since these models are inherently “creative” and therefore non-deterministic by default. Non-determinism means that each generation can produce slightly different results, such as different wording of a sentence or different pixels in a generated image. In highly regulated industries, such as financial services, this is particularly challenging as AI models must be explainable (both internally and to regulators), and organizations must be able to prove their outputs are correct. Nobody wants “hallucination” in a banking or payments transaction.

AI practitioners in these industries can mitigate the risks of non-determinism with strategies like grounding, which instruct models to base their outputs on authoritative reference information, such as real-time data from enterprise systems or policy documents. This context can be provided directly in the prompt itself or using techniques like retrieval augmented generation (RAG). Furthermore, you can adjust model parameters, such as temperature, to instruct the model to generate answers that are more factual.

It’s imperative to bring everyone in the organization together to learn the techniques being leveraged for model controls and governance and set up a positive feedback loop between teams.

Importantly, no matter which technical approaches are used to gain confidence in model behavior, a robust testing and validation system should be developed alongside the AI system itself. Evaluation is critical during development and should continue after production deployment, so that stakeholders and the AI product team can verify that the system is performing as expected over time. Logging is key — and don’t forget that AI tools can summarize logs and perform outlier detection to spot issues like model drift. It’s imperative to bring everyone in the organization together to learn the techniques being leveraged for model controls and governance and set up a positive feedback loop between teams.

Financial services institutions are increasingly aware of the significant role they can play in addressing climate change. As allocators of capital through their lending and investment portfolios, they direct financial resources for corporate development and operations in the wider economy. 

This capital allocation responsibility balances growth opportunities with risk assessments to optimize risk-adjusted returns. Identifying, analyzing, reporting, and monitoring climate risks associated with physical hazards, such as wildfires and water scarcity, is becoming an essential element of portfolio risk management.

Implementing a cloud-native portfolio climate risk analytics system

To help quantify these climate risks, this design pattern includes cloud-native building blocks that financial services institutions can use to implement a portfolio climate risk analytics system in their own environment. This pattern includes a sample dataset from RS Metrics and leverages several Google Cloud products, such as BigQuery, Data Studio, Vertex AI Workbench, and Cloud Run. The technical architecture is shown below.

Please refer to the source code repository for this pattern to get started, and read through the rest of this post to dig deeper into the underlying geospatial technology and business use cases in portfolio management. You can use the Terraform code provided in the repository to deploy the sample datasets and application components in your selected Google Cloud Project. The README has step-by-step instructions.

After deploying the technical assets, we recommend performing the following steps to get more familiar with the pattern’s technical capabilities:

• Review the example Data Studio dashboard to get familiar with the dataset and portfolio risk analytics (see screenshot below)
• Explore the included R Shiny app, deployed with Cloud Run, for more in-depth analytics
• Visit Vertex AI Workbench and walk through the exploratory data analysis provided in the included Python-based Jupyter notebook
• Drop into BigQuery to directly query the sample data for this pattern

The importance of granular objective data

Assessing exposure to climate risks under various climate change scenarios can involve combining geospatial layers, expertise in climate models, and using information about company operations. Depending on where they are located, companies’ physical assets – like their manufacturing facilities or office buildings – can be susceptible to varying types of climate risk. A facility located in a desert will likely experience greater water stress, and a plant located near sea level will have a larger risk of coastal flooding.

Asset-level physical climate risk analysis

Google Cloud partner RS Metrics offers two data products that cover a broad set of investable public equities: ESGSignals® and AssetTracker®. These products include 50 transition and physical climate risk metrics such as biodiversity, greenhouse gas (GHG) emissions, water stress, land usage, and physical climate risks. As an introduction to these concepts, we’ll first describe two key physical risks: water stress risk and fire risk.

Water Stress Risk

Water stress occurs when an asset’s demand for water exceeds the amount of water available for that asset, resulting in higher water costs or in extreme cases, complete loss of water supply. This can negatively impact the unit economics of the asset, or even result in the asset being shut down. According to a 2020 report from CDP, 357 surveyed companies disclosed a combined $301 billion in potential financial impact of water risks.

When investors don’t have asset location data, they use industry average water intensity and basin level water risk to estimate water stress risk, as described in a 2020 report by Ceres. However, ESGSignals® allows a more granular approach, integrating meteorological and hydrological variables at the basin and sub-basin levels, drought severity, evapotranspiration, and surface water availability for millions of individual assets.

As an example, let’s look at mining, a very water-intensive industry. One mining asset, the Cerro Colorado copper mine in Chile, produced 71,700 metric tons of copper in 2019, according to an open dataset published by Chile’s Ministry of Mining. ESGSignals® identifies this mining asset as having significant water stress, resulting in a water risk score of 75 out of 100. For assets like these, reducing water consumption via efficiency improvements and the use of desalinated seawater will not only save precious water resources for nearby communities, but also reduce operating costs over time.

Fire Risk

Wildfires have caused significant damage in recent years. For example, economists estimated that the 2019-2020 Australian bushfire season caused approximately A$103 billion in property damage and economic losses. Such wildfires pose safety and operational risk for all kinds of commercial operations located in Australia.

ESGSignals® fire risk score is calculated by combining historical fire events, proximity, and intensity of fire with company asset locations (AssetTracker®). Based on ESGSignals® assessments, the majority of mining assets located in Australia have medium to high exposure to fire risk.

Incorporating asset-level climate risk analytics into portfolio management 

Now that we have an understanding of the mechanics of asset-level climate risk, let’s focus on how portfolio managers could incorporate these analytics into their portfolio management processes, including portfolio selection, portfolio monitoring, and company engagement.

Portfolio selection

Portfolio selection can involve various investment tools. In screening, the portfolio manager sets up filtering criteria to select companies for inclusion in, or exclusion from, the portfolio. Asset-level climate risk scores can be included in these screening criteria, along with other financial or non-financial factors. 

For example, a portfolio manager could search for companies whose average asset-level water stress score is less than 30. This would result in an investment portfolio that has an overall lower risk from water stress than a given benchmark index (see figure below).

Portfolio monitoring

For portfolio monitoring, it’s important to first establish a baseline of physical climate risk for existing holdings within the portfolio. A periodic reporting process that looks for changes in water stress, wildfire, or other physical climate risk metrics can then be created. Any material changes in risk scores would trigger a more detailed analysis to determine the next best action, such as rebalancing the portfolio to meet the target risk profile.

Portfolio engagement

Some portfolio managers engage with companies held in their portfolios, either through shareholder initiatives or by meeting with corporate investor relations teams. For these investors, it’s important to clearly identify the assets with significant exposure to climate risks. 

To focus on the locations with the highest opportunity for impact, a portfolio manager could sort the millions of AssetTracker locations by water stress or fire risk score, and engage with companies near the top of these ranked lists. Highlighting mitigation opportunities for these most at-risk assets would be an effective engagement prioritization strategy.

Expanding beyond portfolio management

Applying an asset-level approach to physical climate risk analytics can be helpful beyond the use cases in portfolio management presented above. For example, risk managers in commercial banking could use this methodology to quantify lending risk during underwriting and ongoing loan valuation. Insurance companies could also use these techniques to improve risk assessment and pricing decisions for both new and existing policyholders.

To enable further insights, additional geospatial datasets can be blended with those used in this pattern via BigQuery’s geospatial analytics capabilities. Location information in these datasets, such as points or polygons encoded in a GEOGRAPHY data type, allow them to be combined together with spatial JOINs. For example, a risk analyst could join AssetTracker data with BigQuery public data, such as population information for states, counties, congressional districts, or zip codes available in the Census Bureau US Boundaries dataset.

A cloud-based data environment can help enterprises manage these and other sustainability analytics workflows. Infosys, a Google Cloud partner, provides blueprints and digital data intelligence assets to accelerate the realization of sustainability goals in a secure data collaboration space to connect, collect, correlate information assets such as RS Metrics geospatial data, enterprise data, and digital data to activate ESG intelligence within and across the financial value chain.

Curious to learn more? 

To learn more from RS Metrics about analyzing granular asset-level risk metrics with ESGSignals®, you can review their recent and upcoming webinars, or connect directly with them here.

To learn more about sustainability services from Infosys, reach out to the Infosys Sustainability team here. If you’d like a demo of the Infosys ESG Intelligence Cloud solution for Google Cloud, contact the Infosys Data, Analytics & AI team here.

To learn more about the latest strategies and tools that can help solve the tough challenges of climate change across industries, view the sessions on demand from our recent Google Cloud Sustainability Summit.

^{Special thanks to contributors
The authors would like to thank these Infosys collaborators: Manojkumar Nagdev, Rushiraj Pradeep Jaiswal, Padmaja Vaidyanathan, Anandakumar Kayamboo, Vinod Menon, and Rajan Padmanabhan. We would also like to thank Rashmi Bomiriya, Desi Stoeva, Connie Yaneva, and Randhika H from RS Metrics, and Arun Santhanagopalan, Shane Glass and David Sabater Dinter from Google.}

^{Disclaimer
The information contained on this website is meant for the purposes of information only and is not intended to be investment, legal, tax or other advice, nor is it intended to be relied upon in making an investment or other decision. All content is provided with the understanding that the authors and publishers are not providing advice on legal, economic, investment or other professional issues and services.}

In December, I was fortunate to be a guest on Cloudreach’s Cloudbusting Podcast, where I spoke with Jez and Dave about Google’s sustainability initiatives. This was my first time being a guest on a podcast, and it was a very fun experience! Sustainability has been increasingly important theme for me, and there was a lot to talk about.

We started with some definitions of terms that people use when talking about sustainability: decarbonization, which I defined as transitioning energy production from fossil fuel sources to carbon-free & renewable sources, and ESG, which refers to focusing on environmental, social, and governance aspects of companies and is a common term in the world of sustainable finance, especially for investors.

Next we talked about Google’s approach to sustainability, including all of the highlights from Google’s sustainability journey — carbon neutral in 2007, 100% renewable energy matching since 2017, and the most recent goal of moving to 24×7 carbon free energy by 2030. On this last point, I mentioned that this goal is a moonshot, which it definitely is! But it’s also a series of roofshots, since we’re incrementally making progress towards this goal. In fact, as we move each cloud region to a higher percentage of carbon-free energy, organizations can move their cloud workloads accordingly to minimize footprint. Note that this data is also available in machine-readable formats in GitHub and BigQuery.

Then we discussed COP26, the United Nations Climate Change Conference in Glasgow. This was happening when this podcast was recorded on November 5th, 2021. I reviewed the main goals for the conference including keeping 1.5 alive, protecting communities and habitat, and mobilizing finance. We didn’t get too deep into these points, and there was just so much happening at COP that it was frankly a little overwhelming. I’m still catching up several months later!

It will become part of your job as an engineer to understand your cloud carbon footprint

From there we discussed the sustainability announcements from Google Cloud Next ’21, our annual conference. This included Carbon Footprint, where you can see the gross carbon emissions associated with the electricity generation of your cloud usage by product, project, and region. The hosts made a great connection here between “Sustainability Ops” and FinOps, which is super insightful, and allowed me to make a prediction: “It will become part of your job as an engineer to understand your cloud carbon footprint”. You heard it here! (Probably not first, others have probably made this point too, but still…) Check out this Next session to dig into how this tool works in more detail:

We also touched on Active Assist Recommender, a tool that uses machine learning to find cloud projects that are likely abandoned. The idea is that by deleting these projects, you can reduce carbon emissions.

The last major announcement we covered was the introduction of the Google Earth Engine commercial use pilot. I have to admit that I had not heard of Earth Engine before joining Google, but it’s a super impressive product that has been around for more than 10 years and has become very useful in academia for science involving earth observation. It combines petabytes of satellite imagery that is stitched together and maintained by Google with a geospatial analytics processing engine.

Previously, only research and nonprofit organizations have been able to use Earth Engine (by the terms of service), so the new cool thing being piloted is a commercial license that would allow for-profit companies and public sector (governments) to use it as well. One of the primary reasons for commercial organizations to use Earth Engine is to further their sustainability efforts with geospatial approaches. As an example, I talked about how Unilever is using Earth Engine to help ensure a deforestation-free supply chain for commodities like palm oil.

Finally, we wrapped up the podcast with a bit of fun in the form of “ten second recommendations” which is a regular feature of this podcast. My contribution was Ready Player Two. I particularly enjoyed the battle of the Seven Princes!

Hopefully this will be the first of many podcast experiences in the future. If you enjoyed it and would like to discuss more about sustainability, or would like to invite me to join your podcast, please connect with me on Twitter or LinkedIn!

(Featured photo by me, taken in a Google office.)

Technology is playing a vital role in this effort. I have a background in financial services and fintech, and work in Google Cloud’s CTO Office with technology leaders across financial services and other industries to transform sustainability in their organizations. Here are three technology areas that I think are vital for sustainable finance, both to manage climate-related risks and to enable new climate finance opportunities.

Understanding localized climate risks

COP26 leaders are calling on companies to be transparent about the risks inherent in climate change. Every business will be affected by a warming planet, and it’s critical to understand precisely how these events will affect company operations, employees’ lives, and entire value chains. As BlackRock’s CEO Larry Fink has stated, climate risk is investment risk.

However, physical risks from climate change are not evenly distributed around the world. Some areas will see increased floods, and others will face droughts or wildfires. Rising sea levels will affect coastal areas and severe winter storms will affect places with historically milder winters.

Geospatial analytics and modeling has the potential to give us a much more accurate view of these localized climate-related physical risks. The Task Force on Climate-Related Financial Disclosures (TCFD) recommends factoring the geographic location of the organization’s value chain (both upstream and downstream), as well as the organization’s own assets, into any climate change scenario analysis efforts. Using tools like Google Earth Engine and BigQuery, it’s possible to create a localized, precise understanding of climate-related physical risks for each location in the value chain.

Using AI to improve climate disclosure

Another key finance priority for COP26 is improving the quantity, quality and comparability of climate-related disclosures by companies, so that investors and regulators can better understand climate risks and opportunities. To date, most of this disclosure has been published as human-readable documents. One example is Google’s Environmental Report. These documents are very helpful if you’re in a position to deeply study an individual company, but what if you want a broad view of sustainability progress across markets?

This is where AI comes in. Specifically, natural language machine learning approaches can be applied to classifying and extracting structured information from sustainability disclosures. The TCFD recently used an AI approach to quantify the level of sustainability disclosure across 1,701 companies from 69 countries. A team of experts manually labeled text passages in a sample of 150 documents, assigning a “yes” or “no” value to each review question, such as “does the company disclose Scope 1, Scope 2, and, if appropriate, Scope 3 greenhouse gas emissions?”. Then they trained a model with these labels and used it to predict yes/no responses for the entire corpus. The resulting analytics were used by the task force to assess the state of adoption and level of disclosures.

Investors, regulators, and companies can all benefit from adopting natural language AI techniques to better understand sustainability documents, whether those are from suppliers, investors, customers, or even competitors. Tools like Google Document AI and Vertex AutoML can accelerate AI adoption in these organizations.

Collaborative sustainability data sharing

COP26 is demonstrating the power of collaboration to solve climate challenges. It’s clear that individual organizations cannot go at this alone; everyone must work together! In the financial world, companies need to share information with investors, suppliers need to share data with companies, and everyone needs to share data with regulators and central banks.

In an enterprise setting, sharing data can be difficult, involving legacy tools, slow cycle times, and one-way processes. What if sharing climate and sustainability data were as simple as sharing photos or documents in the cloud? Cloud-native tools like Analytics Hub, built on BigQuery, can definitely help here.

As a bonus, thanks to Google’s renewable energy matching, building these solutions on Google Cloud today means the net operational greenhouse gas emissions associated with your application is zero. And over time, the gross carbon footprint of these workloads will decrease as we make progress towards our goal of running on carbon-free energy around the clock by 2030.

This article originally appeared in the Google Cloud Blog.

What’s driving this shift in mentality across every industry? It’s simple: Consumers are no longer willing to support companies that don’t prioritize sustainability. According to a survey conducted by IBM, the COVID-19 pandemic has elevated consumers’ focus on sustainability and their willingness to pay out of their own pockets for a sustainable future. In tandem, federal action on climate change is increasing, with the U.S. rejoining the Paris Climate Agreement and a recent executive order on climate commitments.

Over the past few years, we have seen an uptick in organizations setting long-term sustainability goals. However, CEOs and chief sustainability officers typically forecast these goals, and they are often long term and aspirational — leaving the near and midterm implementation of ESG programs to operations and technology teams.

CTOs are a crucial part of the planning process, and in fact, can be the secret weapon to help their organization supercharge their ESG targets. Below are a few immediate steps that CTOs and technology leaders can take to achieve sustainability and make an ethical impact.

Reducing environmental impact

As more businesses digitize and more consumers use devices and cloud services, the energy needed by data centers continues to rise. In fact, data centers account for an estimated 1% of worldwide electricity usage. However, a forecast from IDC shows that the continued adoption of cloud computing could prevent the emission of more than 1 billion metric tons of carbon dioxide from 2021 through 2024.

Make compute workloads more efficient: First, it’s important to understand the links between computing, power consumption and greenhouse gas emissions from fossil fuels. Making your app and compute workloads more efficient will reduce costs and energy requirements, thus reducing the carbon footprint of those workloads. In the cloud, tools like compute instance auto scaling and sizing recommendations make sure you’re not running too many or overprovisioned cloud VMs based on demand. You can also move to serverless computing, which does much of this scaling work automatically.

Deploy compute workloads in regions with lower carbon intensity: Until recently, choosing cloud regions meant considering factors like cost and latency to end users. But carbon is another factor worth considering. While the compute capabilities of regions are similar, their carbon intensities typically vary. Some regions have access to more carbon-free energy production than others, and consequently the carbon intensity for each region is different.

So, choosing a cloud region with lower carbon intensity is often the simplest and most impactful step you can take. Alistair Scott, co-founder and CTO of cloud infrastructure startup Infracost, underscores this sentiment: “Engineers want to do the right thing and reduce waste, and I think cloud providers can help with that. The key is to provide information in workflow, so the people who are responsible for infraprovisioning can weigh the CO2 impact versus other factors such as cost and data residency before they deploy.”

Another step is to estimate your specific workload’s carbon footprint using open-source software like Cloud Carbon Footprint, a project sponsored by ThoughtWorks. Etsy has open-sourced a similar tool called Cloud Jewels that estimates energy consumption based on cloud usage information. This is helping them track progress toward their target of reducing their energy intensity by 25% by 2025.

Make social impact

Beyond reducing environmental impact, CTOs and technology leaders can have significant, direct and meaningful social impact.

Include societal benefits in the design of your products: As a CTO or technology founder, you can help ensure that societal benefits are prioritized in your product roadmaps. For example, if you’re a fintech CTO, you can add product features to expand access to credit in underserved populations. Startups like LoanWell are on a mission to increase access to capital for those typically left out of the financial system and make the loan origination process more efficient and equitable.

When thinking about product design, a product needs to be as useful and effective as it is sustainable. By thinking about sustainability and societal impact as a core element of product innovation, there is an opportunity to differentiate yourself in socially beneficial ways. For example, Lush has been a pioneer of package-free solutions, and launched Lush Lens — a virtual package app leveraging cameras on mobile phones and AI to overlay product information. The company hit 2 million scans in its efforts to tackle the beauty industry’s excessive use of (plastic) packaging.

Responsible AI practices should be ingrained in the culture to avoid social harms: Machine learning and artificial intelligence have become central to the advanced, personalized digital experiences everyone is accustomed to — from product and content recommendations to spam filtering, trend forecasting and other “smart” behaviors.

It is therefore critical to incorporate responsible AI practices, so benefits from AI and ML can be realized by your entire user base and that inadvertent harm can be avoided. Start by establishing clear principles for working with AI responsibly, and translate those principles into processes and procedures. Think about AI responsibility reviews the same way you think about code reviews, automated testing and UX design. As a technical leader or founder, you get to establish what the process is.

Impact governance

Promoting governance does not stop with the board and CEO; CTOs play an important role, too.

Create a diverse and inclusive technology team: Compared to individual decision-makers, diverse teams make better decisions 87% of the time. Additionally, Gartner research found that in a diverse workforce, performance improves by 12% and intent to stay by 20%.

It is important to reinforce and demonstrate why diversity, equity and inclusion is important within a technology team. One way you can do this is by using data to inform your DEI efforts. You can establish a voluntary internal program to collect demographics, including gender, race and ethnicity, and this data will provide a baseline for identifying diversity gaps and measuring improvements. Consider going further by baking these improvements into your employee performance process, such as objectives and key results (OKRs). Make everyone accountable from the start, not just HR.

These are just a few of the ways CTOs and technology leaders can contribute to ESG progress in their companies. The first step, however, is to recognize the many ways you as a technology leader can make an impact from day one.

This article originally appeared in TechCrunch.

How do you hire a software engineer these days? It had been a while since I hired a technologist, so I felt a little rusty posting a position for the Observer.

The last time I hired technical talent, it was for an early-stage, recently funded, financial technology startup. It was a little different. The goal then was to hire as many top programmers as quickly as possible to build a completely new product. The clock was ticking, and we were burning our venture funding.

Before that, I hired software developers within a large financial software company, S&P Capital IQ. We already had a big technology team, with more than a thousand people in a bunch of offices around the world, in fact. So there were tons of resources, like a dedicated HR staff and an established campus-recruiting machine at top colleges. It was more about wading through resumes then.

Now I’m looking for the first full-stack software engineer at the Observer. We’re pretty small, but we’re not a startup. (A fact that my significant other reminds me of frequently when I code on weekends… Some habits die hard, I suppose.)

Aside: for those that haven’t heard the term “full stack” before, for web software it basically means writing code that runs in the browser as well as on the server. It’s a fancy way of saying you need a generalist rather than a specialist. On small development teams, generalists are the way to go.

Who do I know?

I started, as many hiring managers do, by looking at my LinkedIn connections and trying figure out who I knew that might be a good fit. Of course, it’s great to hire someone you’ve worked with before and who you know and respect. But this can be complicated. It’s largely a timing thing: you have to catch people when they are between jobs or thinking of leaving. If you’re off by a month or even a few weeks, you’ve lost them to another company.

If you get lucky and find someone in your network who’s looking to move, it’s still not a guarantee. They might have a different skill set than is needed for the role, or they might be interested in working remotely when that’s not what you had in mind, or you might envision their personality clashing with others on the team. It’s a bad idea to force it.

Of course, you can always try to poach—to go after people in your network who are ensconced in a gig, happily coding away, headphones on, in some open working space with a cool open floor-plan. But that’s a dangerous game. Maybe the poachee’s company employs other people you used to work with? Maybe you used to work with said poachee’s boss? Poach, and suddenly you’ve made a career enemy who will return the favor down the road. Maybe there are other capable engineers at the same company who will feel jilted and badmouth you or your team at the bar? Engineers are humans too (well, most of us) so we gossip. Better to stick with the seekers, I think.

Just post it

Which brings us to the traditional route, at least since the late 90’s: posting a job online. Usually there’s a baseline site of some kind, like a career portal that your company hosts or utilizes. Ours is LinkedIn, so we started there.

Pete Soderling, founder of Hakka Labs, an engineering culture platform, talks about the art of crafting a good job posting for engineers. There are many crappy job descriptions for tech positions out there. Unless you’re looking for a boring employee, try not to write a boring one.

However, just posting that non-boring job description on LinkedIn is not enough. Many talented engineers tend to view LinkedIn with trepidation. It’s something you kind of have to have, and it can be a great way to connect with other professionals. But most tech people get a lot of recruiter trolling from LinkedIn. A former coworker of mine used to have a standardized blow-off response in a little text file stashed on his computer that he would copy & paste for recruiter InMails. Engineers are always trying to automate, after all.

It works the other way, too: when you post a tech job listing on LinkedIn, it’s often recruiters who find you first. “I have five top developers with experience in ___, ____, and ____.” They fill in the blanks with the first three tech acronyms quoted in your posting, even if those make no sense together or are super generic. “Can I send you their resumes?” A previous manager gave me this sage advice: never say yes to this. They don’t actually have five top developers who fit the job. It can be a great thing to work with recruiters, but get a relationship going with them before that job posting. Otherwise they know you’re desperate. Which, of course, you are.

To the apps

So where to go next? This being 2015, to the apps! There are a few that look interesting. We’ve posted our job on Switch, which is kind of like Tinder for jobs. You install the app, post your job description, and then start swiping left for candidates who don’t fit and swiping right for ones who look interesting. Job candidates also swipe left and right, and if they swipe right, Swipe sends you both an intro email. [Editor’s note 2021: Switch is no longer active.]

Another Tinder-for-Jobs option that’s been floating in my Twitter stream is Poachable. This app is similar to Switch, but it relies on anonymity a little more, and has a more involved process. For example, you need to schedule a phone meeting with Poachable staff before they will post your job on the system, and they provide customizable screening questions to help you narrow down the candidate pool. Employers pay for credits that allow them to ask candidates who have expressed interest to reveal their identities. It’s a little creepy, but then again so is online dating, and this is pretty similar. [Editor’s note 2021: Poachable also seems to be shut down.]

Finally, I’ve heard good things about Hired.com. I have yet to try it out, but the premise is similar: it’s a marketplace for job seekers and employers to connect. Hired promises a $2,000 hiring bonus to job seekers, and wants 15 percent of the first year base salary from the employer for a successful hire. That’s lower than what you’d typically pay for a recruiter, and two grand is nothing to sneeze at, so I bet they are pretty successful. [Editor’s note 2021: They were! Hired is still around. And as I recall, we wound up finding a great engineer via Hired a few months after this was posted.]

Meetups

Another strategy is to go to tech meetups, or even sponsor them. There are a million in New York, and many of them have a ton of high quality, mostly free or inexpensive content. Some of them are amenable to recruiting, especially by engineering managers, but some are actively hostile to all recruiting.

Note: in the original piece, there was a call to action at the end: apply to said software engineering job at Observer. Obviously that doesn’t make sense any more. But if you’re interested in talking with me about hiring, team-building, and similar, get in touch!

This Spatial Data Science Conference is a gathering of data science, advanced analytics, and GIS experts from across the globe. The geospatial community has experienced significant growth as businesses and organizations utilize location-based analytics to adapt to the new era we are living in.

I would expect this trend to continue as earth observation and machine learning tools become more powerful. According to a March 2021 research report, the global geospatial analytics market size was valued at $58.35 billion in 2019, and is projected to reach $158.84 billion by 2027, growing at a CAGR of 14.2% from 2020 to 2027. For comparison, in 2020, AWS had $45.37 billion in revenue. So the geospatial market is big.

Geospatial analytics are crucial for sustainability, because looking at the surface of the Earth can tell us a lot about the effects of climate change, especially using satellite imagery. For example, here is a Google Earth Timelapse of the Columbia Glacier retreat in Alaska:

Here are some of the specific technologies and products that Nathan and I talked about during our talk:

• BigQuery CARTO Spatial Extension
• CARTO for React using the open source project deck.gl
• Google Cloud product discovery solutions for retail including Search for Retail
• Google Earth Engine, used to monitor global deforestation by Global Forest Watch
• Google Cloud Vertex AI

Doing this talk gave me an opportunity to reflect on the technology aspects of geospatial analytics as well. Things have progressed quite a lot since 2014, when I last worked extensively in the space. At that time, I was CTO of an early-stage startup that focused on retail store location analysis for investors. This phrase didn’t exist at the time, but today you would have called us an “alternative data” provider. Here’s how we described this functionality in our sales deck:

To build these store location analytics, we collected (scraped) retail store locations from retailer websites and geocoded the street addresses to capture latitude and longitude for each store. Then we calculated “trade areas”, or shapes around each store representing the area where customers were mostly likely traveling from to shop. We used a few different methods to do this, including “simple rings” (5-mile radius, etc.) and “travel time” (e.g. 10-minute driving time, also called isochrones). Ultimately each trade area become a geometric Polygon, an array of points, that we stored in the database.

After creating the trade area polygons, we computed demographics like total population and median household income for space within each trade area using US Census data. Finally, we aggregated these stats for each retail chain and provided tools to compare different retail chains, and understand how these metrics changed over time with new store openings and store closings.

Our primary toolset at the time included ESRI, MongoDB, PostGIS, and Python. If I had to do this now, I would definitely start with BigQuery and more data science friendly tools like CARTO. It would be much simpler!

At any rate, please enjoy the talk, and feel free to contact me on Twitter or LinkedIn if you’re interested in any of this geospatial sustainability stuff.