Rohan Biswas

AI Engineer

https://www.opentalent.in/rohan-biswas

AI Engineer with less than a year in Python & Data Science

Indian Institute of Science (IISc)

Key Strengths

Strong practical experience in building and deploying end-to-end ML systems, including a production-grade RAG pipeline and a full-stack forecasting system.
Proficiency in key AI/ML frameworks and tools such as PyTorch, TensorFlow, LangChain, HuggingFace Transformers, and FastAPI.
Demonstrated ability to translate research prototypes into production-ready code, including containerization with Docker and deployment on platforms like HuggingFace Spaces.
Experience with advanced ML concepts like agentic RAG, LSTM for time-series forecasting, PCA for dimensionality reduction, and rigorous validation frameworks (out-of-sample, data leakage, overfitting).
Solid understanding of data engineering aspects, including data ingestion, feature engineering, and database interaction (PostgreSQL, ChromaDB, vector databases).
Strong problem-solving and analytical skills, evidenced by identifying failure modes, root cause analysis, and improving simulation fidelity in research internships.

Cultural & Operational Fit

Cultural Fit Analysis

The candidate's project diversity, ranging from financial forecasting to agentic RAG systems for epidemic modeling and adaptive PCA, indicates a broad interest and adaptability, which is a strong cultural fit for dynamic AI engineering environments. Their experience in both academic research (IISc) and practical deployment (HuggingFace Spaces) shows a blend of theoretical understanding and real-world application. The competitive FAST-SF Fellowship and Best Undergraduate Project award highlight a drive for excellence and a proactive learning attitude, aligning well with innovation-driven cultures.

Soft Skills & Operational Fit

The candidate demonstrates strong problem-solving, analytical, and communication skills through their detailed project descriptions and research internship experience. Their ability to document failure modes, translate complex outputs into actionable summaries, and work on multi-agent orchestration frameworks suggests a structured and collaborative approach to problem-solving. The focus on reproducible workflows and rigorous validation indicates a strong operational fit for roles requiring robust and reliable AI systems.

AI is analyzing your overall score…

Identifying your key strengths…

Evaluating your skill match against the job requirements…

Assessing your cultural and operational fit

Projects

EpiRAG - Production Agentic RAG System

March 1, 2026 – Present

Developed a Python-based hybrid agentic RAG pipeline over 19 epidemic modeling papers (10,700 chunks) using LangChain orchestrating Llama/OpenAI/Qwen debate agents in parallel and OpenAI synthesis, with confidence-based routing (sim-threshold 0.45) between a local ChromaDB vector store and live DuckDuckGo/Tavily web search. Citation enrichment via Semantic Scholar, OpenAlex, PubMed; real-time SSE debate streaming. Containerized and deployed the system as a microservice on HuggingFace Spaces via Docker, with on-startup corpus rebuilding from HF Datasets to enable zero disk storage on server.

Nifty50 Stock Forecasting System - Full Stack ML

December 1, 2025 – Present

Engineered a full-stack LSTM forecasting pipeline in Python: ingested raw market data into PostgreSQL, performed feature engineering on 15+ signals (returns, momentum, volatility, lag features) using pandas, and trained models in TensorFlow. Served predictions via FastAPI with a live Streamlit dashboard; built a rigorous out-of-sample validation framework with systematic diagnosis of data leakage, overfitting, and regime instability across market conditions. Documented each failure mode and root cause end-to-end, from data ingestion through deployment, demonstrating full ownership of the ML lifecycle.

Adaptive PCA Portfolio Analytics

December 1, 2025 – Present

Built rolling PCA factor decomposition pipeline across 9 global equity indices using adaptive 60-day windows to detect regime shifts and non-stationarity in live data. Full backtesting loop: Sharpe ratio tracking (2.4 in-sample), drawdown analysis, and regime-sensitivity diagnostics focused on interpretable output.

View Project

Key Strengths

Strong practical experience in building and deploying end-to-end ML systems, including a production-grade RAG pipeline and a full-stack forecasting system.
Proficiency in key AI/ML frameworks and tools such as PyTorch, TensorFlow, LangChain, HuggingFace Transformers, and FastAPI.
Demonstrated ability to translate research prototypes into production-ready code, including containerization with Docker and deployment on platforms like HuggingFace Spaces.
Experience with advanced ML concepts like agentic RAG, LSTM for time-series forecasting, PCA for dimensionality reduction, and rigorous validation frameworks (out-of-sample, data leakage, overfitting).
Solid understanding of data engineering aspects, including data ingestion, feature engineering, and database interaction (PostgreSQL, ChromaDB, vector databases).
Strong problem-solving and analytical skills, evidenced by identifying failure modes, root cause analysis, and improving simulation fidelity in research internships.

Cultural & Operational Fit

Cultural Fit Analysis

Soft Skills & Operational Fit

Rohan Biswas

Key Strengths

Cultural & Operational Fit

About

Top Skills

Skills

Education

Experience

Projects

Key Strengths

Cultural & Operational Fit