SB
Open to Opportunities

Soumyadeep
Basak

Data Scientist & ML Engineer · B.Tech CS @ IEM Kolkata

SB
Soumyadeep Basak
ML Engineer · Data Scientist
9.08
GPA
6+
Projects
2
Papers
Deep Learning RAG ICPC Regionalist
IIT Ropar Research Intern
WEBEL Data Science Intern
ICPC 2025 Regionalist
Kaggle Silver Badge
ICAA 2026 · ICTIS 2026
Hackathon Top 3 · HackOasis 1.0
Murf-AI Hackathon Winner
Hacktoberfest Top Contributor
IIT Ropar Research Intern
WEBEL Data Science Intern
ICPC 2025 Regionalist
Kaggle Silver Badge
ICAA 2026 · ICTIS 2026
Hackathon Top 3 · HackOasis 1.0
Murf-AI Hackathon Winner
Hacktoberfest Top Contributor

Skills & Technologies

Python95%
TensorFlow / Keras88%
Scikit-Learn & ML90%
LangChain / RAG / LLMs85%
C++ (Competitive Programming)82%
FastAPI / Flask / Streamlit87%
PostgreSQL / MySQL / MongoDB80%
Data Analysis & Visualization89%
AWS Cloud72%
Docker / Git / DevOps78%
PythonC++C TensorFlowKerasScikit-Learn PandasNumPySeaborn NLTKOpenCVLangChain HuggingFaceLlama.cppXGBoost FastAPIFlaskStreamlit PostgreSQLMySQLMongoDB RedisAWSDocker GitJupyterIPFS EthereumStatsmodelRAGAS

Work Experience

Data Science Intern
WEBEL
Apr 2025 – Jul 2025 📍 Kolkata, India
  • Recommender System: Built a personalized hybrid recommender using content-based & collaborative filtering on service records, queried via PostgreSQL.
  • Algorithm Engineering: Built a hybrid similarity approach using demographics, historical behavior, service metadata, co-occurrence patterns, and eligibility constraints to recommend across 400+ public services.
  • Interactive Deployment: Implemented a FastAPI backend supporting real-time recommendations and scheduled model retraining.
Recommender Systems Collaborative Filtering FastAPI PostgreSQL
Research Intern
IIT Ropar
May 2024 – Jul 2024 📍 Ropar, India
  • Insurance Policy Assistant: Engineered an insurance policy assistant RAG, enabling clause lookup and verdict classification.
  • Retrieval Optimization: Optimized retrieval by experimenting with chunking strategies, HyDE, multi-query expansion, and hybrid search.
  • Memory: Integrated per-session memory and dynamic source data updates with tombstone metadata.
RAG LLMs HyDE LangChain

Featured Projects

🤖
Smart CAPTCHA System
Deep Learning · Anomaly Detection · Real-time
A frictionless security solution analyzing real-time user behavior—mouse movements, keystrokes—to silently detect bots. Lowered CAPTCHA challenges by 60% via behavior-based filtering with full analytics dashboard.
PythonTensorFlow Scikit-LearnFlask PostgreSQL
📈
Poly-Market Analyzer
Deep Learning · Time Series · Finance
Resource-efficient stock forecasting using co-movement-based clustering and shared-layer-LSTM with Ticker-aware-LSTM models with custom embeddings to predict Nifty50 trends — one model for many stocks.
PythonTensorFlow StatsmodelStreamlit
🔒
CryptoSecure
Anomaly Detection · Blockchain · Real-time
Platform that assigns safety scores to crypto wallets by combining XGBoost (supervised) and Auto-Encoder (unsupervised) ML models with graph-based anomaly detection, factoring transaction behavior, wallet age, and KYC status.
PythonXGBoost TensorFlowFastAPI Redis
🧠
SecureMind
RAG · Chatbot · Blockchain
Secure, localized RAG chatbot integrating LLMs with document-based Q&A. Supports on-device inference and uses IPFS (InterPlanetary File System) for decentralized storage — works offline or in remote environments.
PythonEthereum HuggingFaceLlama.cpp Streamlit
🔍
LogSense
Log Analysis · RAG · Anomaly Detection
Pipeline to parse large-scale Linux logs, extract structured features (PID, process, templates), detect anomalies using statistical and embedding methods, and use a context-aware RAG for automatic incident diagnosis.
LangChainScikit-Learn FastAPIRAGAS
🖱
Session Impostor Detector
Behavioral Biometrics · Security
Session-level impostor detection using mouse-based behavioral biometrics. Analyzes movement patterns, velocity, and click behavior to distinguish genuine users from impostors — accepted at ICTIS 2026.
PythonScikit-Learn PandasFeature Engineering

Education

🎓
B.Tech in Computer Science
Institute of Engineering and Management
Kolkata, India · Aug 2022 – May 2026
GPA: 9.08
📚
CBSE — Class XII & X
Hariyana Vidya Mandir
Kolkata, India · 2009 – 2022
XII: 91.66% X: 94%

Languages

🇬🇧 English Fluent
🇮🇳 Bengali Native
🇮🇳 Hindi Proficient

Achievements

🏆

ICPC 2025 Regionalist

Competitive programming · Specialist on Codeforces (1406) · 3★ on CodeChef (1720)

🥈

Kaggle Silver Badge

Top 5% on AIMO competition · Top 20% at Bamboo Summer challenge

🚀

Hackathon Top 3

Secured top 3 ranking at HackOasis 1.0 and Murf-AI Hackathon

🌍

Hacktoberfest Top Contributor

Named Top Contributor in the open-source project nigeria-crime-trends

Publications

C.1
Stocks in Sync: Cluster-Aware Deep Learning for Multi-Stock Forecasting in Financial Market
ICAA 2026 · Kolkata, India
Conference Paper
C.2
Session-Level Impostor Detection Using Mouse-Based Behavioral Biometrics
ICTIS 2026 · Thailand
Conference Paper

Get In Touch

📱
📍
Location
Kolkata, India
Status
Open to Opportunities
SB