LinkedIn · Portfolio · Email

$ whoami
vignesh_kumar_u — quant researcher · ai engineer · builder

$ cat status.json
{
  "role": "WorldQuant BRAIN Research Consultant",
  "regions": ["USA", "EUR", "KOR", "ASI", "IND", "CHN"],
  "graduation": "B.Tech, AI & Data Science — May 2026",
  "shipping": "PaperPulse — open-source research digest tool",
  "open_to": ["Quant Research", "Data Science", "AI/ML Engineering"]
}

I work at the intersection of quantitative research, machine learning, and financial technology — alpha signal research on live trading platforms, LSTM-based anomaly detection, semantic search pipelines, and fraud analytics, mostly validated against real or high-fidelity simulated data.

Currently seeking entry-level roles in quantitative research, data science, and AI/ML engineering.

WorldQuant BRAIN — Quantitative Alpha Research Python Statistical Modeling Time-Series Analysis Ongoing alpha research on WorldQuant's simulation platform. 60+ submitted models across global equity universes (USA, EUR, KOR, ASI, IND, CHN), validated against Sharpe, Fitness, and self-correlation thresholds. Diagnosed in-sample vs. out-of-sample degradation as an independent overfitting signal.

PaperPulse — Research Paper Trust-Scoring Engine Python NLP Quant Signal Design Open-source tool that scores and digests research papers using quant-derived trust signals, applying alpha-validation rigor from live trading research to academic literature triage.

RakshaSetu — Smart Tourist Safety Monitoring System LSTM Autoencoders Geo-Fencing Ethereum Solidity First-author research system for real-time tourist safety monitoring, accepted and presented at IEEE CONIT 2026. LSTM autoencoder for GPS trajectory anomaly detection (seq2seq reconstruction), 92% anomaly detection accuracy, sub-2-second alert response. Simulated dataset of 250,000+ GPS samples across 15,000 registered users. Geo-fencing via ray-casting with a blockchain-based digital identity layer.

LEARNING-AID — Semantic OER Recommendation System FastAPI Whisper spaCy KeyBERT FAISS Qdrant Real-time speech-to-text (Whisper / faster-whisper) with segment-level confidence scoring. Multi-method NLP concept extraction (spaCy NER + KeyBERT). Semantic vector search benchmarked FAISS vs. Qdrant (150–250ms) over 1,000+ resources. Fully open-source, no proprietary APIs.

Hercules Finance AI — Financial Planning for Gig Workers Python FastAPI TypeScript React Gemini Top 30 nationally at SEBI Global FinTech Festival 2025. Income forecasting for irregular earners (RMSE/MAPE validated), sentiment-based financial distress detection, crisis planning, and fraud simulation. Multilingual (English, Hindi, Marathi) with UPI simulation.

Financial Crime Detection Pipeline Python Scikit-learn NetworkX Tableau Graph-based suspicious network detection (strongly connected components, betweenness centrality) on 1M+ transactions. Risk scoring engine with automated flag generation and an interactive Tableau dashboard with drill-down reporting.

DeepWork AI — Productivity Desktop App TypeScript Python Electron MCP Top 50 globally at Logitech Dev Studio Hackathon 2026. Hardware-triggered focus session timer via Logitech MX Creative Console, browser workspace capture/restore (Chrome, Edge, Brave), and a notification triage dashboard (TF-IDF + logistic regression).

Languages: Python, R, TypeScript, SQL AI/ML: Scikit-learn, PyTorch, FAISS, Qdrant, spaCy, Whisper, Gemini Backend & Infra: FastAPI, React, Docker, PostgreSQL, Redis, Kafka Visualization: Tableau, Power BI, Streamlit

RakshaSetu: Multi-purpose Smart Tourist Safety Monitoring and Incident Response System based on AI, Geo-Fencing, and Blockchain Technology First Author · IEEE CONIT 2026 · Accepted & Presented (Paper ID 227) · Publication pending

AI-Enhanced PGRKAM Employment Analytics Platform IEEE ISAECT 2025 · Published in IEEE Xplore · DOI: 10.1109/ISAECT68904.2025.11318700