AI Engineer passionate about building innovative GenAI and Agentic AI solutions that solve real-world problems.
As the first engineering hire, built 80%+ of Lamatic's core stack. Architected serverless executor processing 1M+ monthly requests, designed AI evaluation framework with LLM-as-a-judge, and built Kubernetes-based ETL pipelines with OAuth systems.
Fine-tuned open-source LLMs (Mistral, LLaMA-2) using LoRA/PEFT and 8-bit quantization on bilingual datasets, improving fluency by 30%. Built scalable LLM Inference Engine with dynamic batching, multi-GPU support, and PagedAttention achieving 106 tokens/sec. Optimized Whisper V3 ASR with ONNX/TensorRT and Triton deployment, reducing latency to 0.1-0.4s.
Built serverless multilingual sentiment analysis system on AWS reducing processing time by 50%. Enhanced NER and text classification models by 30%. Implemented anomaly detection system reducing false positives by 60%.
Developed ML model, converted to ONNX, and deployed with Golang runtime for client-side productivity tracking.
AI-powered content orchestration monorepo that guides you through the full creation pipeline — clarify, outline, write, edit, and publish to Blogger — with autonomous agents handling each stage.
A Model Context Protocol server exposing 35+ tools over the Dev.to (Forem) API v1. Lets AI agents like Claude and Cursor draft, edit, publish, and manage articles programmatically. Supports stdio, HTTP, and Cloudflare Workers transports, with a multi-arch Docker image on GHCR.
End-to-end voice assistant supporting 10+ Indian languages including Hindi and Marathi. Built speech-to-text with Conformer models on Triton, text-to-speech with Fastpitch, and a RAG pipeline using LangChain with embedding and reranker models for knowledge base queries.
Chat with any PDF — an early LLM-powered assistant that lets users ask natural language questions against uploaded documents, with grounded answers retrieved directly from the source.
Explored why memory is the missing infrastructure layer for production AI agents — covering long-term memory implementation, reliable and explainable AI systems at scale, and a live deep dive into Lamatic.ai's agent-building platform.
Presented how Model Context Protocol (MCP) enables LLMs to connect with live data and tools in real-time, breaking free from rigid APIs and static integrations. Included hands-on Python and OpenAI demos.
Discussed why RAG is a system design problem, not just a feature — covering common failure modes, retrieval strategies, evaluation loops, and why prompt engineering alone falls short in production.
For decades, we believed the magic of software was in the instructions — the code. Now, with AI, a quiet revolution is underway: the quality of what you feed an AI matters more than how you build it.
Language was AI's first great breakthrough. But human intelligence was never just about words. Multimodal AI is the ability of a single model to understand and reason across text, images, audio, and video — not separately, but together.
There's a subtle but seismic shift happening in AI right now. For years, AI was reactive — you asked, it answered. That era is ending. Agentic AI doesn't wait. It plans, acts, evaluates, and tries again.
LLMs forget by design. Here's how external memory makes AI agents more personalized, consistent, and useful over time.
The moment prompt tricks fail and engineering finally takes over.
Stop wasting tokens — discover the new data format replacing JSON for LLMs. A 60% efficiency boost inside.
A deep dive into Nemotron 3 Nano, Super, and Ultra - architecture, benchmarks, and why sparse models win on token economics.
Learn how to create a streaming MCP server and connect to it with GPT-4o mini, step-by-step.
Studied core CS subjects including Data Structures, Algorithms, DBMS, OS, and AI. Completed additional coursework in Deep Learning, Data Science, and Machine Learning.
Active member of global AI and developer communities — collaborating, learning, and contributing alongside engineers, researchers, and founders pushing the frontier of AI.
Community of AI engineers and founders building production AI systems.
Network of founders and operators shaping the next wave of startups.
Open-source LLM observability — collaborating on tracing, evals, and prompt management.
Developers building on Workers, AI, and edge infrastructure at global scale.
One of the largest AI learning communities — discussing papers, models, and techniques.
Open-source agentic AI framework community (formerly AutoGen) for multi-agent systems.
Democratizing AI research — curated papers, prompt engineering guides, and applied AI learning.
Always interested in new opportunities, collaborations, and conversations about AI, ML, and building innovative solutions.