I build AI systems that do real work, not just impressive demos.

I'm a Senior Data Scientist at MangoBytes, working at the intersection of AI agents, data engineering, backend systems, and oil & gas domain intelligence. My current obsession is building reliable agentic software that can reason over enterprise data, use tools safely, and actually survive production.

• 🤖 Building multi-agent / tool-using AI systems and AI-native workflows
• 🛢️ Applying AI to oil & gas operations, production, regulatory data, and decision support
• 🧠 Deep into LLMs, agent orchestration, evals, observability, RAG, MCP, and agent security
• 🏗️ Designing production systems across APIs, databases, cloud infrastructure, and data pipelines
• 🔐 Especially interested in making agents secure, permission-aware, auditable, and hard to break
• 🎓 M.Sc. in Artificial Intelligence & Machine Learning, Coimbatore Institute of Technology

Current philosophy: the interesting part of AI isn't making the model smarter — it's engineering the system around the model so it stays useful when reality gets messy.

Also frequently around: LLM APIs · Agent SDKs · MCP · RAG · SQL · REST APIs · Linux · CI/CD · Azure Container Apps · vector / full-text search · data pipelines · observability

AI agents that can be trusted
> clever prompts

production reliability
> perfect demos

clear domain context
> generic AI answers

good system boundaries
> hoping the model behaves

I'm always interested in conversations around AI agents, applied LLM systems, data platforms, developer tooling, agent security, and energy technology.

If you're building something where AI has to interact with the real world and not fall apart, we'll probably have plenty to talk about.