AI ENGINEER · BENGALURU, IN
Two years deep in RAG pipelines, LLM integrations, and the Java & Python services that keep them fast, observable, and in production — not just in a notebook.
I'm an AI Engineer who came up through backend systems first — which means I think about an LLM the way I think about any other dependency: what's the latency budget, what happens when it's slow, and how do I know if it's wrong before a customer does.
At Altolabs, that means owning pieces of an AI-powered platform — retrieval pipelines, LLM integrations, and the Java and Python services that keep them fast and observable in production. Before that, at Mondee, I worked on AI-driven personalization and automation. Each role has sharpened the same instinct: a model is only as good as the system feeding it and the one reading its output.
Off the clock, I run the same discipline on my own hardware — a self-hosted MicroK8s homelab on Raspberry Pi 5 nodes and a Proxmox box, with TrueNAS behind it for storage, hosting Immich for photos, Nextcloud for files, n8n for automation, and Firefly III for personal finance — plus a fair number of 2 a.m. outages I've had to debug myself. It's the same instincts either way — happy to talk through it.
Building and maintaining an AI-powered platform across multiple backend microservices — LLM integration, RAG architecture, and semantic search for accurate, context-aware conversations at production scale.
Built an AI-powered personalization platform delivering intent-driven recommendations and automated workflows using RAG, vector search, and behavioral analysis.
Early hands-on work across backend systems and AI tooling that set the direction for everything since.
Building and scaling an agentic AI platform end to end — LLM integration, RAG architecture, and semantic search to deliver accurate, context-aware conversations, backed by the backend performance work that real production scale demands.
An enterprise-grade AI document intelligence chatbot on Azure for ADIB, giving secure, role-based access to CRM and policy knowledge. RAG workflows and semantic search, built for compliant deployment in a regulated environment — improving agent productivity, response speed, and governance.
An AI-powered travel planning platform delivering end-to-end, intent-driven recommendations and automated booking workflows — RAG, vector search, and behavioral analysis combined into a personalized, scalable travel experience.
A production-like multi-node cluster built around reliability and data persistence — MicroK8s for orchestration across Raspberry Pi 5 nodes and an x86 Proxmox machine, TrueNAS for durable ZFS-backed storage, Cloudflare ingress, full observability, and GitOps for resilient, auditable operations. Built iteratively over a year on a tight budget, with deliberate tradeoffs that favor data safety over uptime. The cluster runs a handful of services I depend on daily: Immich as a self-hosted photo library (with a Python/ffprobe pipeline I built to migrate years of photos out of Google Photos and catch corrupted files along the way), Nextcloud for personal file storage, n8n as the automation backbone — running infrastructure audits and a personal finance pipeline that parses bank SMS alerts and categorizes them automatically — and Firefly III on the receiving end, turning that pipeline into hands-off bookkeeping.
Have something that needs
an AI Engineer who ships?
Let's talk.