I'm a Founding Engineer who joined Lamatic as the first engineering hire and ended up building over 80% of the product from scratch - AI systems, backend infrastructure, and everything in between.

At Lamatic, I architected a serverless executor that now handles 1M+ monthly requests and a deployment engine that cut latency from 2 minutes down to 15 seconds. I also built an AI evaluation framework with LLM-as-a-judge, a multimodal hiring agent that reduced recruiter workload by 70%, and Kubernetes-based ETL pipelines integrated with Drive, S3, and SharePoint. It's been a wild ride of shipping things fast and making them actually work in production.

Before that, at Samespace, I got deep into LLM fine-tuning - working with Mistral and LLaMA-2 using LoRA/PEFT - and built an inference engine hitting 106 tokens/sec with multi-GPU support. I also optimized Whisper V3 ASR models down to 0.1-0.4s latency using ONNX and TensorRT. That's where I really learned what it takes to make AI reliable at scale, not just impressive in a demo.

Outside of work, I write on Medium about things I genuinely find interesting - AI agents, inference optimization, MCP servers. A few of those posts took off, which tells me there's a real appetite for engineers who can explain the hard stuff clearly.

I hold a B.E. in Computer Engineering from the University of Mumbai with a 9.6 GPA. I'm now looking to go further - building something of my own at the intersection of agentic AI and real-world automation