I'm pursuing my MS in Artificial Intelligence student at NTU Singapore , with a BComp in Computer Science with Honours (Highest Distinction). I love building things that sit at the boundary of disciplines: where AI meets biology, where machine learning meets markets and where distributed systems meet real-world constraints.
My work spans LLM applications and RAG pipelines for enterprise intelligence, time-series forecasting for operational decisions, graph ML for cancer drug target discovery, and information retrieval systems over massive document corpora. I'm drawn to problems where the data is messy, the stakes are real, and the solution requires more than off-the-shelf models.
I'm particularly excited by the intersection of AI with computational biology , quantitative finance and AI safety. Outside of code, I'm learning French and Spanish, and trained in Bharatnatyam-which, it turns out, teaches you a lot about precision under pressure.
Python for ML/AI pipelines and research · Java for distributed systems and backend · C/C++ for systems-level and cryptography work · JavaScript for full-stack applications
Gradient boosting · random forests · time-series forecasting · dimensionality reduction (PCA, autoencoders) · SHAP explainability · causal inference · graph-based ranking
Transformers · GNNs (GCN, GAT, GraphSAGE) · CNNs · RNNs · reinforcement learning · transfer learning · fine-tuning · sentence transformers · embedding models
RAG pipelines · semantic retrieval · hybrid search (BM25 + embeddings) · prompt engineering · LLM fine-tuning · regulatory intelligence · multi-agent systems · NER and aspect extraction
Exploratory data analysis · statistical modeling · forecasting dashboards · SHAP/LIME interpretability · interactive visualization · policy-driven insight generation
Full-stack development · JWT authentication · API design · distributed systems · UDP protocols · at-most-once execution semantics · fault-tolerant architectures
AWS (EC2, S3, SageMaker, Lambda) · containerised ML pipelines · CI/CD · Databricks for big data workflows · version control and collaborative development
Relational and NoSQL design · analytical queries · vector databases · hybrid search indexing · full-text search · OLAP
| Domain | Focus |
|---|---|
| 🧬 Computational Biology | Drug target discovery · disease network analysis · ML over biological graphs |
| 📈 Quantitative Finance | LLM-powered alpha signals · regime detection · causal inference in markets |
| 🤖 AI Safety | LLM failure modes · adversarial robustness in high-stakes domains |
| 🔍 Information Retrieval | Hybrid search · semantic ranking · opinion mining at scale |
ML ranking framework using LightGBM, XGBoost, and network-science features to prioritize cancer drug targets across breast and prostate signaling networks. Identified 566 novel target combinations with up to 86% Recall@1 for prostate cancer.
LightGBM XGBoost Graph ML Network Science Python
Low-latency UDP-based distributed banking system with custom binary protocols and at-most-once execution guarantees. Achieved 100% transaction consistency under 50% packet loss via request deduplication and response caching.
Java UDP Distributed Systems Fault Tolerance
AI-powered search over 90K+ developer discussions combining BM25, embedding-based semantic retrieval, and hybrid search on 69K+ indexed documents. Integrated sentiment analysis, NER, aspect extraction, and sarcasm detection.
Elasticsearch Sentence Transformers BM25 NLP Python
Led a team of 8 to optimise pathfinding algorithms for autonomous robot navigation. Integrated turn-radius and obstacle-avoidance heuristics to reduce total route cost by ~25% and improve traversal stability by 30%+.
Algorithms Pathfinding Python Robotics Team Lead