I'm an AI Engineer based in Bangalore, India, focused on building Enterprise-grade AI agent systems that actually work in production. I hold an M.Tech in Artificial Intelligence and Data Science from Devi Ahilya University, Indore.
Right now I'm deep in the weeds of agent architecture — specifically how to make LLM-powered agents more reliable, context-aware, and useful at scale. My work sits at the intersection of production AI systems, enterprise automation, and intelligent agent design.
Currently building AI-powered procurement intelligence at Procureyard — automating the boring, complex parts of enterprise procurement with agents.
- 🏗️ Enterprise AI Agents — Designing and shipping multi-agent systems for real business workflows, not just demos
- 🔍 Agent Context Engineering — Studying how to structure, compress, and maintain context in long-running production agents
- 🧩 Agent Harness & Orchestration — Exploring frameworks like LangGraph, Deer-Flow, and custom orchestration layers to control agent behavior reliably
- 🗣️ LLM Reliability in Production — Prompt engineering, fallback strategies, memory management, and grounding techniques for enterprise deployments
- 🔗 RAG & Knowledge Systems — Building retrieval pipelines that actually retrieve the right thing (Text-to-SQL, knowledge graphs, vector search)
AI / LLM
OpenAI Claude Groq Ollama LangChain LangGraph RAG Vector DBs Prompt Engineering
Languages
Python TypeScript JavaScript SQL PineScript
Backend & Infra
FastAPI Node.js PostgreSQL Redis Docker Google Cloud Platform
Frontend
React Next.js TailwindCSS
Tools
Portkey Perplexity API SQLAlchemy Alembic Selenium
| Project | Description | Stack |
|---|---|---|
| deer-flow | SuperAgent harness — researches, codes, creates with sandboxes, memory & subagents | Python |
| AutoSQLNLP | Natural language to SQL query engine powered by LLMs | Python |
| vanna | Text-to-SQL generation via RAG — chat with your database | Python |
| Text_to_KnowledgeGraph | Convert unstructured text into structured knowledge graphs | Python |
| gpt2_from_scratch | GPT-2 implementation from scratch for deep LLM understanding | Python |
- Advanced agent memory architectures (episodic, semantic, procedural)
- How to reduce hallucination and improve grounding in production agents
- Multi-agent coordination patterns and inter-agent communication
- Scaling agentic workflows for enterprise procurement use cases
- Options trading and algorithmic strategy development
If you're building in the enterprise AI / agent systems space, or just want to talk LLMs, RAG, and production AI — I'm always down for a conversation.
📬 LinkedIn | 🐦 Twitter/X | 💼 Procureyard
"Every problem is solved by yet another indirection."