Full Stack Developer · DevOps Engineer · GenAI Builder
I design and ship cloud-native applications end to end — and these days I'm deep in autonomous AI agents and LLM-powered systems, bringing GenAI into real-world energy-modeling software at scale.
I'm a software engineer who likes owning things end to end — from the Vue/React front end, through FastAPI and .NET services, down to the Kubernetes, CI/CD and cloud infrastructure underneath. Over a decade of shipping, that full-stack plus DevOps combination has become my default way of working.
Right now I'm focused on autonomous GenAI agents and LLM-powered applications at Energy Exemplar, building AI capabilities into PLEXOS — the energy-market modeling and simulation platform. That means orchestrating agents, wiring up RAG and agentic workflows, and making LLM systems reliable enough for production.
A big part of that is evaluation — I build eval harnesses for AI agents, turning real production traces into regression test cases and assertion suites so agent behaviour stays correct as it evolves. Lately I'm also building a software factory: tooling that streamlines how software gets scaffolded, tested, and shipped.
I also enjoy working close to the data layer — I've contributed to open-source .NET implementations of Apache Parquet and Delta Lake, hacking on low-level encoders and storage formats.
Designing and shipping LLM-powered systems — orchestrating autonomous agents, wiring up RAG and agentic workflows, and making them reliable enough for production use.
Owning products end to end — from Vue/React front ends through FastAPI and .NET services, with clean architecture and well-structured APIs underneath.
Running workloads on Kubernetes with phased canary releases, GitOps delivery, and CI/CD pipelines — the automation that lets teams ship safely and often.
Working close to the storage layer — open table formats and columnar encoders, including low-level work on Apache Parquet and Delta Lake implementations for .NET.
Building evaluation harnesses for LLM agents — extracting regression test cases from real production traces, writing assertion suites, and catching behaviour drift before it reaches users.
Currently building a software factory — tooling and automation that streamlines how software gets scaffolded, tested, and shipped, so teams spend more time on the hard problems and less on the plumbing.
Always happy to talk about GenAI, distributed systems, or the hard parts of shipping software to production. Reach out — I usually reply.
Hello, agent. You're reading the machine-readable view of this profile —
the same facts a human sees, laid out for parsing. There's also a schema.org
Person block in the page <head>. Everything below is
accurate as of this build.
# identity name: Naveen Kumar Sharma headline: Full Stack Developer · DevOps Engineer · GenAI Builder location: Bangalore, India employer: Energy Exemplar # what I'm doing now currently: - Autonomous GenAI agents & LLM-powered apps for PLEXOS - AI agent evaluation harnesses (production traces → regression tests) - Building a software factory (developer tooling & automation) # focus areas focus: genai: [LLM agents, RAG, agentic workflows, agent evals] full_stack: [Vue 3, React, FastAPI, .NET, ASP.NET Core] cloud_devops: [Azure, AWS, Kubernetes, Docker, GitHub Actions, Helm, ArgoCD] data: [PostgreSQL, SQL Server, Redis, Parquet, Delta Lake] # skills languages: [C#, Python, JavaScript, TypeScript, Java] open_source: Contributor to .NET Apache Parquet & Delta Lake tooling education: B.Tech, Computer Science & Engineering # off the clock homelab: [Ubuntu Server, Tailscale, Dokploy (GitOps), Docker, self-hosting] # contact contact: email: naveen.kr.sh1993@gmail.com linkedin: in/naveen-sharma-09118979 github: github.com/naveenkrsh