Mumbai, India · B.Tech Computer Science & Engineering (2024–2028)
I’m Yug Shah, an undergraduate Computer Science researcher and builder at SVKM’s Shri Bhagubhai Mafatlal Polytechnic & College of Engineering, Mumbai.
My work focuses on applied machine learning, cybersecurity forensics, quantitative risk modelling, and autonomous multi-agent pipelines. Whether forecasting Antarctic sea-ice drift for Smart India Hackathon, identifying frame-level synthetic artifacts for IIT Delhi eRaksha, or heading mathematical game theory tournaments, I build software that balances mathematical rigor with sensory craft.
An AI-driven decision support system for Antarctic sea-ice forecasting and iceberg trajectory prediction. Built for Smart India Hackathon.
Traditional maritime routing in polar regions relies on lagging meteorological reports and static ice-edge approximations, increasing vessel collision risks. With HimDrishti, we engineered a full-stack decision support platform combining XGBoost anomaly flags with Attention-LSTM time-series forecasting and real-time A* pathfinding over FastAPI to deliver dynamic, collision-free transit corridors.
Hover over titles to scramble letters, or click “Grow in Garden” to plant their seeds into the botanical canvas.
Polar navigation decision-support architecture engineered for Antarctic sea-ice concentration forecasting and multi-day iceberg drift trajectory prediction. Combines spatio-temporal Attention-LSTM networks with risk-weighted A* vessel pathfinding to guide research icebreakers through shifting cryospheric corridors safely.
Agentic deepfake detection and multimodal media authentication engine built for law-enforcement and digital forensics. Dissects spectral audio anomalies, facial landmark phase-coherence, and EXIF provenance chains to detect synthetic generation in real time.
Full-stack quantitative portfolio risk analyzer and market simulator. Computes institutional risk metrics including Value at Risk (VaR), Sharpe & Sortino ratios, max drawdown regimes, and Monte Carlo multi-asset stress test simulations.
Threat-surface reduction and data-segregation research initiative conducted with CERT-In to educate citizens, students, and SMBs on proactive cyber hygiene, phishing countermeasures, and credential privacy defenses.
End-to-end orchestration pipeline powering the FPP Careers digital publication — automating semantic script structuring, SEO metadata generation, thumbnail prompt synthesis, and scheduled distribution.
Snapshots from hackathons, midnight math derivations, and building late into the night. Hover to lift and inspect any polaroid!
Spearheaded The Matrix ’26 (flagship quantitative and game theory championship) and Project Chanakya (18-hour inter-college industry hackathon and venture pitch arena) for 600+ students.
Awarded Best Speaker in Finance serving as CFO representing BlackRock — debating institutional portfolio allocation, sovereign debt risk profiles, and macroeconomic liquidity policy under scrutiny.
Completed CS50 Cybersecurity (Harvard / edX) with distinction and formulated the Google Student Ambassador product concept pitch introducing the Pixel Y Developer Edition.
Yug’s approach to quantitative risk modeling was impressively rigorous for someone at the undergraduate level. His StockPulse project showed a rare blend of mathematical depth and engineering polish.
The CyberRaksha deepfake detection system was one of the most technically ambitious submissions at eRaksha. Yug’s multi-agent forensics pipeline genuinely surprised our panel of judges from IIT Delhi.
Leading SANKHYA while concurrently building complex ML systems is no small feat. Yug runs The Matrix championship with the precision of someone who actually understands game theory.
Open for research collaborations in AI/ML, cyber forensics, quantitative finance, and full-stack engineering.
