AI & ML Engineer
I take models from data to deployment: training, fine-tuning, RAG pipelines, and the infrastructure that keeps them running.
Technology is at its best when it disappears. I build machine learning systems that learn from real data - aligning models with intent, predictions with measurable quality, and infrastructure with the products they serve.
The dataset decides the ceiling. I spend the time on data quality, labeling, and evaluation before reaching for a bigger model or a cleverer pipeline.
Local transformer inference on consumer Intel silicon via OpenVINO before reaching for a datacenter. Sub-24ms tokens, no GPU required.
A notebook is not a system. Streaming inference, caching, observability, and CI gates turn experiments into products that hold up under real load.
The work behind the headline: tap each one for the specifics.
End-to-end AI and ML systems. From data pipelines and model training to multi-agent orchestration and production deployment.
Autonomous academic research with multi-agent fact-checking - three agents (Researcher, Analyst, Critic) collaborate in a feedback loop streamed live over SSE, with a React 19 knowledge graph frontend.
Python / LangGraph / FastAPI / React 19 / LLM Agents / Ollama / OpenAI / PyMuPDF
View sourceMulti-agent LLM code review with deterministic cross-checking - 23 review categories, every finding validated against the actual repository before it ships, shipped as the acr CLI with a full-screen Textual terminal dashboard.
Python / LLM Agents / Typer CLI / Textual TUI / AST Analysis / Taint Analysis / SARIF / OpenAI / Anthropic / Gemini
View sourceProfessional timeline and academic credentials in artificial intelligence and machine learning.
Leading AI system design and implementation for enterprise-scale document intelligence, including RAG pipeline architecture, model fine-tuning, and production ML infrastructure.
Built the AI backend for a personalized adaptive learning platform, including student performance modeling and an LLM-based tutor using retrieval-augmented generation.
Developed AI automation pipelines for lead intelligence and data enrichment, and built ML/NLP models integrated into production workflows.
Specialized in Artificial Intelligence and Machine Learning. Completed thesis on Generative AI.
Explored the core concepts and algorithms of modern AI, including search algorithms, knowledge representation, machine learning, neural networks, natural language processing, and reinforcement learning - all implemented in Python.
Comprehensive introduction to machine learning covering clustering, dimensionality reduction, supervised learning algorithms, data analysis and manipulation, model evaluation, and statistical learning theory.
“Working with Sourav was smooth and highly professional. He delivered quality quickly and elevated the entire product experience.”
“Sourav brought both technical depth and design clarity. The final website was modern, intuitive, and beyond expectations.”
References available on request. Happy to connect you directly.