Software engineering · Security · Applied AI
I build software, assess technical risk, and help teams make informed decisions.
My experience spans backend systems for finance and healthcare, AI and identity security at Nokia, and geospatial research at NRC Canada. Along the way, I’ve led engineers, worked directly with clients, and made technical decisions understandable to the people relying on them.
Professional experience
A few examples of how I connect implementation, evidence, and delivery.
Geospatial Data Scientist Co-op
Sep 2026 — PresentPublic sector researchAt NRC’s Ocean, Coastal and River Engineering Research Centre, I process RADARSAT/SAR imagery in Python for freshwater ice mapping. I’m also building web-based visualizations of river ice and winter road networks so non-specialist federal stakeholders can interpret the analysis.
Security Software Developer Co-op
May — Aug 2026Applied AI & platform securityI designed retrieval and model routing for an internal root-cause-analysis agent. Our team’s solution was judged the most cost-effective submission in a company-wide AI initiative. A separate engineering assistant reduced legacy-code investigation time by 25%+.
I assessed vulnerability impact using product context, upstream fixes, and exploitability evidence. I then built an agent-assisted workflow for releases with thousands of findings, recording evidence and decisions in Jira. I also implemented upgrade and rollback for a Keycloak identity platform.
The AI work connected engineering context with retrieval and cost-aware model routing. The separate assistant brought together Jira, Confluence, code, and documentation to support investigations.
I performed vulnerability assessments manually before automating repetitive steps. For the identity platform, I tested in-service upgrade, configuration changes, and recovery across Kubernetes and Helm environments.
Senior Software Engineer
Dec 2021 — Aug 2025Financial services & healthcareI led three engineers delivering backend services for Intuit’s QuickBooks Finance Agent and served as a direct technical contact across three client engagements. My work also included APIs and AWS data workflows for Connxus HIE and Harvard Medical School, connecting engineering decisions to client requirements and release quality.
Built Python REST APIs, FastAPI services, and AWS workflows, including HealthLake, NLP, and UMLS integrations. I also served as the primary DBA point of contact and delivered chatbot services handling approximately 10K calls/hour.
The QuickBooks Finance Agent work supported a product designed to serve more than one million customers.
Projects & tools
Independent tools and team research, with source code you can inspect.
Verification for AI-assisted code changes.
Runs repository-defined checks and inspects changes for weakened tests, protected-file edits, and incomplete verification. CI policies come from a trusted base commit, so a change cannot quietly rewrite its own requirements.
Focus: verification design, evidence, and CI integration.
Making clinical terminology easier to work with.
A Python client for UMLS REST APIs, built for healthcare and clinical data workflows. Published on PyPI and listed in the National Library of Medicine’s UMLS community resources.
Focus: API integration and reusable developer tooling.
Lease-clause assessments with source citations.
Retrieves relevant Canadian statutes with FAISS and uses GPT-4o mini to produce cited assessments. A scheduled workflow refreshes the statute index, connecting the generated explanation to its supporting material.
Focus: retrieval, traceability, and evidence maintenance.
Evaluating efficiency alongside fairness.
A Montreal emergency-services simulator built with two teammates. Reinforcement-learning rebalancing reduced the fairness gap by 12.7% against a heuristic baseline in the simulation, with reproducible evaluation artifacts.
Focus: operational tradeoffs and measurable comparisons.
Also built: LLM Evidence Summarizer — summaries linked to sentence-level evidence.
All repositoriesUpstream contributions
Contributions to identity and access management, with the implementation and review history available in each pull request.
Status checked September 2026 · Links show the latest activity.
Talks & community
Technical storytelling, mentoring, and learning with other developers.
Through Google Cloud community videos, I’ve reached 8K+ students. I’ve also trained 200+ undergraduates through technical sessions and mock interviews, and mentored new engineers at Persistent Systems.
Google Cloud APAC
Our story as the winning team in Google Cloud’s Next Big Thing Hackathon.
Watch on YouTubePersistent Systems
Sharing experience and onboarding insights with engineers starting their careers.
Watch on YouTubePosts from Persistent Systems and Google Cloud.
Persistent’s post about how my Semicolons hackathon experience carried into the Google Cloud competition.
Read on LinkedInGoogle Cloud’s #HumBanayenge spotlight on Team Zacharias and building a solution together under time pressure.
Read on LinkedInA Google Cloud feature on my experience in the winning team and how it shaped my development as an engineer.
Read on LinkedInAbout me
I enjoy work that needs both technical depth and sound judgment: understanding a problem, building a solution, and being able to explain why it makes sense.
Client delivery taught me to think about the people implementing and maintaining a system. Security work sharpened how I assess evidence and document decisions. My current research adds another perspective: making analysis useful to people who may never look at the code.
Contact
I’m always interested in thoughtful conversations about technology, security, research, and the problems teams are trying to solve.
