Fullstack Software Engineer
From databases and cloud infrastructure up to the interface people actually touch - I like turning messy real-world problems into systems that just work.
Hi, I'm Preetam Raj, a fullstack software engineer who loves turning messy real-world problems into systems that just work. I studied Mechanical Engineering with a minor in Computer Science and Engineering at IIT Mandi, and somewhere along the way fell for the craft of building software end to end - from database and infrastructure up to the interface people actually touch.
These days I'm a Software Engineer II in energy-tech / cleantech, where I own a platform that powers smart metering, automated billing, and bill audit for commercial and industrial customers - designing the backend and data pipelines, shipping the frontend dashboards, and running the whole thing on GCP. I like being the person who takes an idea from architecture to production and keeps it alive.
I work AI-native day to day - Claude Code and Codex are a regular part of how I ship - and on the side I build llm-kit, an open-source Python toolkit (on PyPI) that gives one typed interface across Anthropic, Gemini, and OpenAI.
Outside of code, you'll usually find me singing or strumming a guitar - music is where I find my balance. Originally from Patna, now based in Bangalore, I'm always chasing that mix of solid engineering and a little creative spark.
An open-source Python toolkit I build and publish on PyPI - one typed, async interface across Anthropic, Gemini, OpenAI, and Bedrock, with Pydantic-structured outputs and multimodal file handling so you never rewrite code per provider. It grew out of my day-to-day LLM work and now stands on its own.
Led the full move of our entire stack across clouds - working with a team GCP brought in, I drove the migration of databases, backend services, load balancers, the S3 data lake to Cloud Storage, and a WordPress site off Lightsail onto a Compute Engine VM, keeping the platform live throughout. Leading that effort taught me how every layer fits together, and how to steer a team toward a shared finish line.
A Python / FastAPI platform I architected from the ground up - decoupled domains, clean thin-controller → service → repository layering, and the discipline to keep 50K lines maintainable as a team of one. I care a lot about code that stays easy to change.
A real-time pipeline on Kafka and TimescaleDB hypertables that chews through high-volume time-series data and feeds everything downstream - alerting, billing, dashboards. I love the puzzle of making data flow reliably at scale.
A model-agnostic extraction system I designed around a Strategy pattern - pluggable schemas, prompts, and validate-with-retry per format - that turns messy unstructured documents into clean, audited data across dozens of formats. It runs on Anthropic Claude and Google Gemini interchangeably, so I can pick the best model per job. This is where I get to mix good engineering with AI.
An engine that generates monthly PDF invoices (including merged multi-facility ones), takes payments via Razorpay, and delivers over WhatsApp and email - collapsing a slow manual process from days to hours. I enjoy killing tedious work with software.
A service that spots anomalies in live sensor data (current imbalance, power-factor drift) and fires targeted WhatsApp/Slack alerts, so people can act on problems in near real time.
A Playwright automation agent that logs into regional utility portals and pulls documents on its own - because nobody should be doing that by hand.
The infrastructure it all runs on today: GCP end to end - Cloud Run, GKE, Cloud SQL, WordPress on a VM, Cloud Storage, migrations, auth, and leader-elected cron - all operated by me.
The frontend too: Next.js 15 / React 19 dashboards (TanStack Query/Table, Redux Toolkit, Clerk) that make all of the above actually usable.
Have a role or a project in mind? I'm happy to talk.
