Gopal Krishna Tewari

Built a multi-modal ML pipeline to predict protein phase separation using 200-dimensional Word2Vec sequence embeddings fused with biophysical aggregates. Designed a dual-branch hybrid CNN that reached 96% accuracy and 0.98 AUC, and ran high-throughput inference on 6,200+ unannotated proteins.

Engineered an analytical data pipeline to parse 120K+ unstructured records and extract heavy-tailed frequency and combinatorial scaling metrics. Modeled heavy-tailed distributions with MLE and KS optimization and validated Menzerath-Altmann laws with strong convergence (R ≥ 0.995).

Developed a Django backend to automate migration of 150+ programming problems with data integrity, atomicity, and efficient rollback for PostgreSQL and Azure Storage. Improved performance using Redis caching, reducing API data retrieval latency by 75%+, and reduced end-to-end migration time by 84%.

Optimized a mobile application using Expo (React Native), cutting tree diameter analysis time by 95%+ to enable faster carbon credit assessment. Implemented an OpenCV-based Python backend to extract depth-adjusted bounding boxes and compute tree diameters with 82% accuracy, eliminating the need for LiDAR sensors.

Provided detailed explanations and solutions for computer science topics including DSA and OOPs to improve student understanding. Achieved 71% Content Feedback and completed two technical rounds.

Activities and societies: Cracked JEE Mains (99.15%ile) and JEE Advanced (AIR 7398). Finalist in Google STEP Program 2K24 and Adobe India Hackathon 2k25; Fresher runner up at DevHaven Hackathon 2K24.

Pursuing a Bachelor of Technology in Computer Science and Applied Mathematics at IIIT Delhi (2023–2027).