Prachurya Adhikary

I’m a Data Engineer who builds dependable data pipelines and turns complex data into clear, actionable products. I care about scalability, correctness, and observability—so data systems stay trustworthy as they grow.

In my research work at Elder Laboratories–Center for Vision Research, I designed and iterated on 5 MCMC-based models to generate synthetic contour data for downstream computer vision tasks. I implemented an optimized intersection-checking algorithm that eliminated self-intersections (100%) across 80,000+ published synthetic samples. I also improved contour generation efficiency, raising the MCMC acceptance rate from 23% to 56%, and helped secure $20,000 CAD in research funding tied to UN Sustainable Development Goals.

During my MTECH Data Engineer (Remote Co-op) at Sanofi, I migrated Excel-based reporting to AWS Athena-backed SQL views in Power BI, improving dashboard refresh speed by ~5x and removing manual data preparation overhead. I strengthened pipeline observability with structured logging, automated retries, and alerting, reducing debugging time by ~50%. I audited legacy ETL pipelines, resolved 6 critical aggregation errors, and collaborated with data scientists, analysts, and product stakeholders to deliver backend features in a Django system that streamlined batch processing and reporting workflows.

My projects reflect the same end-to-end mindset: engineering the pipeline, designing the interfaces, and delivering insights. In the Bangladesh Inflation Monitor, I built a fault-tolerant ETL consolidating 60+ Excel reports into a PostgreSQL warehouse, orchestrated with Airflow scheduling and two-tier caching; I also exposed 6 FastAPI REST endpoints and delivered a Streamlit analytics dashboard for CPI, wage growth, and real income trends.