I build practical AI systems that move from idea to production fast.

• Website: https://www.craftsmanlabs.net/
• Documentation: https://docs.simpleagents.craftsmanlabs.net/
• SimpleFlow SDK Docs: http://docs.simpleflow-sdk.craftsmanlabs.net/

I'm a Chennai-based AI/ML software engineer and indie builder working under the CraftsmanLabs brand. I focus on agentic AI systems, multimodal LLM applications, memory architectures, and production-grade developer tooling.

One-line background: I started shipping and leading technical communities early, then scaled into full-time AI engineering across IDP, NLP, and agent systems.

• I was a core engineer for this HR-focused project: https://news.microsoft.com/source/asia/features/a-digital-colleague-how-chemist-warehouse-and-insurgence-ai-are-rewriting-the-hr-playbook/

AI is most powerful when we stop treating models as black boxes and start designing systems around them.

I'm exploring the space where:

• human intent defines direction,
• agents execute complex workflows,
• and memory plus orchestration make outputs reliable enough for real-world use.

My focus is simple: vibe-code agents to production without giving up control.

That means combining:

• frontier models for capability,
• SLMs for cost and performance on domain tasks,
• and model sovereignty so teams can decide where intelligence runs.

I'm especially interested in building with open and controllable ecosystems, including LFM and Qwen-class models, while keeping interfaces provider-agnostic.

These packages are published and available to use today.

Package	Version	Downloads
simple-agents-py
simple-agents-node
simpleflow-sdk (PyPI)
simpleflow-sdk (npm)
simple-agent-type
simple-agents-cache
simple-agents-core
simple-agents-ffi
simple-agents-healing

• AI architecture blueprints for enterprise systems and agent platforms
• RL and reasoning methods for LLM workflows (including GRPO and recursive approaches)
• Multi-agent collaboration patterns with sparse communication and tool routing
• RAG preprocessing and context-optimization strategies for better retrieval quality
• Applied AI pipelines in finance, supply chain, healthcare, and scientific discovery

I actively build, fork, and contribute around the agentic ecosystem, including infrastructure, model serving, MCP-based tooling, and production workflow patterns.

If you're building in agent orchestration, memory systems, model routing/proxy layers, or domain-specific AI products, I'd love to collaborate.

Open an issue, start a discussion, or reach out through any repo above.