Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I'm a Boston University computer engineering graduate (ML concentration, math-statistics minor) who builds production-grade agentic AI systems. My work spans a multi-agent biomedical research platform at EMD Serono, agent deployment on IBM's open-source Kagenti platform, and a 7-agent commodities trading system of my own. I'm strongest where AI meets reliability: orchestration, governance, and the evaluation layers that make LLM systems trustworthy enough to depend on. I've spent three years as a quant developer, and I pair deep technical work with the ability to explain it to anyone in the room. Based in Boston.…I'm a Boston University computer engineering graduate (ML concentration, math-statistics minor) who builds production-grade agentic AI systems. My work spans a multi-agent biomedical research platform at EMD Serono, agent deployment on IBM's open-source Kagenti platform, and a 7-agent commodities trading system of my own. I'm strongest where AI meets reliability: orchestration, governance, and the evaluation layers that make LLM systems trustworthy enough to depend on. I've spent three years as a quant developer, and I pair deep technical work with the ability to explain it to anyone in the room. Based in Boston.

Moumit Bhattacharjee

AI Engineer, Data Scientist, Full Stack Developer, +2





I'm a Boston University computer engineering graduate (ML concentration, math-statistics minor) who builds production-grade agentic AI systems. My work spans a multi-agent biomedical research platform at EMD Serono, agent deployment on IBM's open-source Kagenti platform, and a 7-agent commodities trading system of my own. I'm strongest where AI meets reliability: orchestration, governance, and the evaluation layers that make LLM systems trustworthy enough to depend on. I've spent three years as a quant developer, and I pair deep technical work with the ability to explain it to anyone in the room. Based in Boston.…I'm a Boston University computer engineering graduate (ML concentration, math-statistics minor) who builds production-grade agentic AI systems. My work spans a multi-agent biomedical research platform at EMD Serono, agent deployment on IBM's open-source Kagenti platform, and a 7-agent commodities trading system of my own. I'm strongest where AI meets reliability: orchestration, governance, and the evaluation layers that make LLM systems trustworthy enough to depend on. I've spent three years as a quant developer, and I pair deep technical work with the ability to explain it to anyone in the room. Based in Boston.

Available to hire

I’m a Boston University computer engineering graduate (ML concentration, math-statistics minor) who builds production-grade agentic AI systems. My work spans a multi-agent biomedical research platform at EMD Serono, agent deployment on IBM’s open-source Kagenti platform, and a 7-agent commodities trading system of my own. I’m strongest where AI meets reliability: orchestration, governance, and the evaluation layers that make LLM systems trustworthy enough to depend on. I’ve spent three years as a quant developer, and I pair deep technical work with the ability to explain it to anyone in the room. Based in Boston.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Work Experience

Multi-Agentic Commodities Trading System at Personal Project

January 1, 2026 - Present

Architected an event-driven, 7-agent AWS system (Lambda, EC2, Terraform) trading crude-oil (USO) and natural-gas (UNG) ETFs; a coordinator ran a 60-second loop over market-regime detection, 4 parallel strategies, and MCP-served LLM news sentiment (Groq, Gemini fallback). Engineered the real-time data and risk stack: Finnhub WebSocket and Alpaca feeds on AsyncIO/Redis, plus a 7-layer safety system (position sizing, portfolio-heat caps, stop-loss, T+1 settlement, and circuit breakers with a kill switch); 89/89 tests passing.

Cloud Computing Engineer, Internship at IBM Kagenti.ai

September 1, 2025 - December 1, 2025

Deploy AI agents on Kagenti, an open-source cloud-native agent platform: build Docker images, author Kubernetes manifests and Helm values through the Kagenti Operator, and write kubectl/port-forward runbooks for dev-cluster rollouts. Build framework-neutral agents for retrieval-QA, data-enrichment, and form-filling use cases, defining MCP tool schemas/adapters and A2A planner policies, and ship containerized demos with cross-platform (macOS + Windows) docs. Hit acceptance goals of 3 use cases and 4 recorded demos, including a secure enterprise Slack research agent gated by the platform's SPIRE/Keycloak identity and least-privilege token exchange; contribute issues and PRs to the public repo.

AI Engineer, Internship at EMD Serono (Merck KGaA)

September 1, 2025 - May 1, 2026

Architect a dual-orchestration layer (LangGraph agent planner over an MCP routing engine) converting noisy biomedical sources into 18-section, citation-backed competitive-intelligence reports. Orchestrate retrieval across 9 specialized MCP servers (PubChem, Open Targets, STRING-db, PubMed/medRxiv), adding a 384-dim sentence-transformer memory and live verification of every NCT/PMID/DOI citation against ClinicalTrials.gov, NCBI, and doi.org. Own the governance and evaluation layer (IBM Context Forge gateway): per-user/per-MCP rate limiting, PII/PHI scrubbing, audited source attribution, and per-tool latency/success metrics, plus a gold-set hallucination check for tracking report quality.

High-Performance Iterative FFT with Financial Application at Personal Project

August 1, 2025 - Present

Implemented a radix-2 in-place FFT in C with OpenMP (O(N log N)); validated vs NumPy and scaled to 2^24; built Python tools to FFT log-returns (AAPL/SPY) and interpret spectra as volatility/short-term “rhythm.”

Undergraduate Research Lead - Computational Psycholinguistics and LLM R&D at Boston University Psycholinguistics Lab

May 1, 2025 - December 1, 2025

Led a study on culturally adaptive LLMs: dataset design, RAG with culture-specific retrieval, QLoRA fine-tuning; built a conversation-level evaluation harness (human ratings) for cultural appropriateness, hallucinations, and safety. Coordinated IRB/ethics and data curation; baselined Llama-3 / Mistral / GPT-4o with ablations on prompt style and retrieval scope; first-author manuscript in preparation.

Remote Software Engineer – LLM Training at Outlier

May 1, 2024 - September 1, 2024

Built evaluation frameworks for tool-use and parameter validation, and authored code completions/prompt patterns for coding tasks, improving demo-readiness and failure analysis across internal test sets.

Senior Quant Developer at Boston University Finance & Investment Club

January 1, 2023 - Present

Applied Random Matrix Theory and Modern Portfolio Theory to the club's $1,000,000; constructed the Efficient Frontier and optimized risk/return with Python-driven Monte Carlo sims; presented methods to faculty/industry.