Apu Bhowmick

Contributed to a Python prototype combining signal rules, LLM reasoning and evidence checks on synthetic banking data, including a FastAPI interface, SQLite storage and pytest tests.

Designed and programmed a Dutch-language regulatory question-answering prototype using OpenAI embeddings and PostgreSQL/pgvector retrieval over approximately 1,800 passages.

Designed and programmed a physics-guided fault-localisation prototype for an airfield-lighting circuit, ranking candidate fault zones from PLC telemetry and communication-error indicators.

Developed a mask-based classification pipeline for catheter tip position classification from chest X-rays, including skeleton-based tip localisation and comparison of machine learning models.

Contributed to a five-person project using VITO data; the team combined XGBoost, Viterbi smoothing and a Python Shiny dashboard for geological review.

Modelled 4,625 repeated measurements from 185 pears using random-slope mixed models in R, comparing climate/species growth trends with adjusted contrasts and exploring power and sample size.

Contributed statistical analysis in an eight-person course team: data cleaning, multiple regression, confidence intervals, residual diagnostics and VIF checks for soil carbon associations.

Pursuing a Master of Statistics and Data Science with a Data Science specialisation, expected to complete in 2026.

Completed a Bachelor of Science in Computer Science and Engineering from 2018 to 2023.