Jitam Barman

Jitam Barman

A Bit About Me

I am a Dual Degree (BTech + MTech) student at IIT Kharagpur, specializing in Metallurgical and Materials Engineering with a minor in Computer Science and Engineering. My interests lie at the intersection of AI, machine learning, and Materials Engineering.

Education

IIT Kharagpur — BTech & MTech Dual Degree, Metallurgical & Materials Engineering 2023 – Present
Read my blog → My Resources →
Resume Experience Skills Projects Resources

Study Resources

Experience

SNN-based Quadratic Programming

— University of Oulu, Finland
May 2026 - Present
  • Deriving a correspondence between Poisson spike-train noise in an SNN solver and Gaussian uncertainty propagation
  • Formulating a Chance-Constrained QP over alloy composition for multi-property steel design, targeting <5% constraint violation
  • Benchmarking the proposed SNN-CCQP framework against classical convex solvers, evaluating the optimality gap

Dev & Analyst Intern

— Zlice Pvt Ltd
Jun 2026 - Aug 2026
  • Developed backend order-management APIs on a Supabase database, supporting the complete order lifecycle for 2500+ users.
  • Created a geospatial dashboard with Leaflet to plot order coordinates across canteens and delivery zones for a clear view of demand.
  • Built a retention-analytics layer exposing a 53% non-returning cohort and segmented it across 23 halls to drive the re-engagement roadmap.

Skills

Languages
Python C C++ JavaScript SQL HTML / CSS MATLAB
ML / AI
PyTorch TensorFlow Scikit-learn Pandas NumPy SciPy Matplotlib NLP RAG Hugging Face
Web / App
React.js React Native Expo Streamlit Bootstrap Leaflet
Tools & Platforms
Git / GitHub Docker SQLite Bash LaTeX PostHog Neo4j

Projects

VEDAS

A tag-anchored Industrial Knowledge Graph on Neo4j, fusing graph and vector retrieval — a resolution cascade lifted tag accuracy from 50% to 100%, and hybrid retrieval lifted answer coverage by 24 points. Includes a LangGraph RCA agent running evidence-grounded 5-Why analysis and flagging real PESO/OISD compliance gaps.

Neo4j LangGraph Knowledge Graph RAG
View on GitHub

CascadeGuard

A spatio-temporal GNN rail digital twin that predicts delays and re-routes passengers

PyTorch Geometric FastAPI React Native SimPy
View on GitHub

FinRAG

A real-time RAG engine for financial analysis — ingests live news and stock-price streams via Pathway's differential dataflow with zero re-indexing, using two-stage retrieval (bi-encoder recall, FlashRank cross-encoder reranking) for precise, data-backed answers.

Pathway Streamlit SQLite LLM
View on GitHub

Multithreaded MD Simulator

An N-body molecular dynamics simulator in C++ — cell-list decomposition cuts complexity from O(N²) to O(N), and a custom jthread pool delivers a 4.5x speedup

C++ Multithreading Physics Simulation
View on GitHub

Microstructural Crack Propagation Prediction

A* search over a Voronoi-tessellated microstructure with a provably admissible heuristic, validated via a 10,000-trial Monte Carlo reliability analysis.

Python A* Search Monte Carlo
View on GitHub

Explore KGP

A gamified campus exploration app with a "fog of war" mechanic over a 23,456-tile geohash map of IIT KGP

React Native Expo Geolocation
View on GitHub

Position of Responsibility

Head - Sponsorship

— COMPOSIT, IIT Kharagpur
Nov 2024 – Apr 2026
  • Managed and led a sponsorship team of 20 members to secure ₹10 Lakh+ in funds across 2025–26 cycles from 15+ corporate partners and alumni

Secretary

— Azad Hall of Residence
Oct 2024 – Apr 2025
  • Coordinated 4 General Championship events and 2 Intra-Hall initiatives; contributed to the Gold-winning Illumination 2024

Activities

Made with the assistance of Claude Opus