Adrika Ghosh

About

Real-time data doesn't wait for you to be certain. Neither do the models built on it. That's the space I work in — I'm an MSc Data Science candidate at Christ (Deemed to be University), Lavasa, with a BSc in Computer Science from Calcutta University, working at the intersection of data, finance, and intelligent systems. I like projects that go beyond theory. Under Dr. Ramesh Kumar Ayyasamy at Universiti Tunku Abdul Rahman, Malaysia, I built an ML-driven decision-support dashboard forecasting stock trends and constructing efficient portfolios — benchmarking Random Forest, XGBoost, and LSTM models, and applying Modern Portfolio Theory with risk analytics benchmarked against NIFTY 50 and SENSEX. In a separate quant finance project, I optimized a 2-asset portfolio via Sharpe-ratio maximization, delivering a 15% improvement in risk-adjusted returns over an equal-weight baseline. That same instinct — for finding signal in noisy, high-stakes data — took me into mental health. I architected an analytics pipeline on a 10,000-participant clinical dataset, engineered preprocessing workflows for heavy missingness and extreme outliers across multimodal biological signals (HRV, cortisol, fMRI), and benchmarked ML models to a peak F1-score of 0.798 predicting neuro-behavioral outcomes. It's this work that led me to my current research question: whether data-driven systems can truly interpret complex psychological states like trauma and depression, or whether some gaps in human understanding are ones no model can close. I'm looking for opportunities as a Data Scientist — to keep building systems that make sense of complexity, wherever it shows up.

Company history

1. 
   Universiti Tunku Abdul Rahman (UTAR)Data Analyst2026-06 – 2026-08 · 2 mo
2. 
   VOPAISGTM Engineer2026-07 – present · 1 mo

Skills

No playbooks indexed for this profile yet.