About
Engineer at the intersection of scale, AI, and security
I'm a software engineer with 4+ years of experience designing cloud-native backend services and production AI systems for finance, healthcare, and telecom security.
At Nokia, I work on security software — building a lightweight In-Service Upgrade and rollback system (35% lower resource usage, 40% faster upgrades), assessing vulnerabilities across product components, and creating agentic-AI workflow automations that cut triage and investigation effort by 30–40% across security engineering teams.
Before that, I spent nearly four years at Persistent Systems as a Senior Software Engineer, where I led a team of three delivering backend services for Intuit's QuickBooks Finance Agent (1M+ customers), cut high-traffic API response times by 66%, and built NLP and ML pipelines for healthcare clients including Harvard Medical School and Connxus HIE.
Outside of work, I contribute upstream to Keycloak (IAM), publish open-source Python tooling featured by the National Library of Medicine, and have mentored 8K+ students through Google Cloud developer community programs.
Career
Experience & Education
Security Software Developer Co-op
Nokia
- Built a lightweight In-Service Upgrade (ISU) & rollback system — 35% less upgrade-time resource usage, 40% faster execution, 50% faster rollback recovery.
- Shipped AI workflow automations for security teams, cutting repetitive triage and investigation effort by 30–40%.
- Developed an AI assistant connecting JIRA, Confluence, codebases, and docs — reducing legacy-code investigation time by 25%+.
- Assessed vulnerabilities across product components: exploitability, affected code paths, and remediation.
- Created reusable agentic AI skills adopted across 5+ security engineering workflows.
MEng, Electrical & Computer Engineering
University of Ottawa
- CGPA 9.60 / 10.00 — focus on design of scalable intelligent systems.
Senior Software Engineer
Persistent Systems
- Led a team of 3 engineers delivering backend services for QuickBooks Finance Agent — 1M+ customers.
- Optimized high-traffic REST endpoints, cutting average response time by 66% for QuickBooks Pro users.
- Delivered FastAPI services connecting React UIs to LLM endpoints for an AI-assist chatbot at ~10k calls/hr.
- Raised automated test coverage to 99%, reducing post-release defects by 70%; improved PostgreSQL performance by 40%.
- Built NLP, UMLS, and ML pipelines for EMR extraction and clinical research, plus AWS HealthLake workflows.
BE, Electronics Engineering · Minor in Computer Science
Shri Ramdeobaba College of Engineering and Management
- CGPA 9.30 / 10.00 · Best Paper Award at ICESC-2020 for traffic-sign detection research.
Selected Work
Projects
Open-source libraries, applied-AI prototypes, and production tooling. Click any card to view the repository.
LeaseCheck — RAG Clause Auditor
Flags illegal or unfair clauses in Canadian residential leases with cited statute snippets. GPT-4o mini + FAISS, function-calling legality checker, Streamlit chat UI, and a scheduled GitHub Action that rebuilds the vector index from updated statutes.
Explainable LLM Evidence Summarizer
Audit-friendly summarization toolkit that outputs bullets with sentence-level evidence IDs for full traceability. Dual cloud/local backends with a CLI, Python API, and Streamlit demo.
UMLS Python Client
Published PyPI package that simplifies integration with UMLS REST APIs through a modular, developer-friendly interface. Featured on the National Library of Medicine's UMLS community resources page.
Community
Open-Source Contributions
Keycloak — Identity & Access Management
Upstream contributions to one of the most widely used open-source IAM platforms: improved OIDC token-introspection client-authentication error handling, Admin Console timestamp display, and HMAC terminology.
umls-python-client — Published Package
Authored and maintain a Python package for the UMLS REST APIs, recognized by the U.S. National Library of Medicine and listed on its official UMLS community resources page.
Toolbox
Skills & Technologies
Languages & Backend
Cloud & Infrastructure
AI & Machine Learning
Security
Datastores & Testing
Tooling & Frontend
Spotlight
Talks, Media & Publications
Traffic-Sign Detection for Driverless Cars
Best Paper Award · ICESC-2020 · BioScience Biotechnology Research Communications
Multimodal Emotion Recognition System
Springer LNEE 992 · ICMEET-2023
UMLS Community Contribution
Featured by the U.S. National Library of Medicine
Contact
Let's build something together
Open to co-op roles (Sep–Dec 2026), research collaborations, and conversations about backend, AI, and security engineering.
Say hello 👋
The fastest way to reach me is email or LinkedIn — I usually reply within a day.