Projects
A collection of production systems, data pipelines, and open-source tools I've built.
Enterprise & Research Projects
Production systems built at Bell Canada, Johns Hopkins, and other organizations
NTS/MS Archway Pipeline@ Bell Canada
End-to-end ETL pipeline integrating 4 enterprise data sources (SmartPath API, Maximo, IPACT, LDAP) into unified Control Plan reporting. 3-tier architecture (Staging → Jarvis → Analysis) processing 150,000+ records in ~20 minutes with schema-versioned loads across DEV/QA/PROD environments.
Duration Calculation Engine@ Bell Canada
Stateful Python algorithm computing 3 distinct duration metrics (d1, d2, d3) measuring agent work cycles. Refactored flawed sequential method into robust two-pass group-by propagation model using pandas shift operations with bfill/ffill propagation. Achieved zero calculation defects in QA validation.
Data Quality Recovery System@ Bell Canada
Full-stack Root Cause Analysis (RCA) diagnosing systemic data integrity drift from misaligned filtration invariant. Executed three staged historical recasts correcting 78,000+ edge-case records, expanding analytical coverage from 1 to 9+ months (+800%) and raising ticket match accuracy to highest level since system inception.
Query Optimization Framework@ Bell Canada
Re-architected critical performance anti-pattern by replacing direct join to 23-million-row live table with pre-aggregated static reference design. Reduced query runtime from 12 minutes to 2 minutes (83% improvement) and stabilized nightly SLA compliance.
Data Consolidation Layer@ Bell Canada
Built transformation logic to consolidate and standardize values from four different source systems using COALESCE waterfall patterns. Fixed 80,000 product discrepancies through proper source table usage and created unified customer name, product, and ticket number fields.
Bioinformatics Platform@ Johns Hopkins
Open-source full-stack bioinformatics platform following SOLID principles and microservices architecture. Features interactive D3.js visualizations, R Shiny dashboards, and real-time WebSocket data streaming. Reduced genomic analysis load times by 83% and increased platform adoption across 100+ global researchers. Private deployment at Johns Hopkins.
Multi-Omics Data Pipeline@ Johns Hopkins
Scalable data processing pipeline integrating 750+ TB of multi-omics Big Data from DISQOVER, ENCODE, PCAWG, PRIDE, and TCGA. Implemented ML models (SVM-RFE, Random Forests) on HPC infrastructure contributing to discovery of 8 novel biomarkers and accelerating validation timelines by 40%. Private deployment at Johns Hopkins.
Personal & Open Source Projects
Side projects, tools, and contributions to the developer community
Microbiome Explorer
Interactive visualization platform for exploring microbial community data with taxonomic profiling, diversity analysis, and comparative metagenomics tools. Built with React frontend and Python backend for processing 16S rRNA sequencing data.
Anomaly Detection System
Automated data quality and anomaly detection pipelines integrating unsupervised ML (K-Means, DBSCAN) and rule-based heuristics. Integrated into ETL and CI/CD workflows to flag real-time anomalies, boosting data integrity by 30% across large datasets.
Portfolio Website
This cyberpunk-inspired portfolio built with Next.js featuring animated canvas backgrounds with floating symbols, custom cursor with lerp interpolation, glassmorphism design system, and responsive layout. Showcases projects, skills, and experience.
Algorithm Visualizer
W.I.P. Interactive web application for visualizing sorting algorithms, pathfinding algorithms, and data structures. Features step-by-step execution, speed controls, and custom input generation for educational purposes.
SQL Query Optimizer Tool
W.I.P. Command-line tool for analyzing SQL query execution plans and suggesting optimizations. Parses EXPLAIN output, identifies common anti-patterns, and recommends index strategies.
Want to Collaborate?
I'm always interested in working on challenging data engineering and platform projects.
