About
The short version
I'm happiest in the guts of a system — working out why something is slow or broken before touching it, then owning the fix end to end.
At Staples I run the FBS Inbound Orders platform: file ingestion from a mess of sources and formats, automated AS400 order creation, event-driven receipt pipelines, and Temporal holding it all together. I joined as a contractor; when our lead left, I already knew the system end to end — so I took the lead role and kept shipping code every day.
Before that: ML research at MITACS and six published papers on optimization and classification. On the side, I built StackUp — a study tracker that's live in production. The long game is to build things of my own.
- 9+
- microservices in production
- 85%
- order-time reduction
- 6
- published papers
- 25+
- citations, lead paper
Work
Featured project — live in production
StackUp2024 — Present
An AI-assisted study tracker for ML/CS self-study, designed, built, and shipped solo.
A full-stack PWA that turns self-study into a system: the Claude API generates context-aware quizzes from what you're learning, spaced repetition resurfaces weak topics, and a built-in Monaco editor keeps practice hands-on. Shipped end to end — auth, analytics, deployment — to prove I can take a product from idea to production alone.
Node.js · Express · Supabase · Vercel · Claude API · PWA
- 01Context-aware quiz generation via the Claude API + prompt engineering
- 02Spaced-repetition scheduling that resurfaces weak topics
- 03In-browser Monaco code editor for hands-on practice
- 04Streak tracking and progress analytics
- 05JWT authentication and secure session handling
Experience
What I do all day
Staples
Jun 2025 — Present · Remote, USA
Software Engineer (Contract) · Team Lead
Leading architecture and delivery of Fulfillment By Staples (FBS) Inbound Orders — an event-driven platform of 9+ microservices spanning order intake, processing, AS400 fulfillment, and receipts.
Sources
Mailbox · SFTP · Blob
Ingestion
JSON · CSV · EDI · XLSX
Orders
Validation · routing
AS400
Spring Boot · <20s
PostgreSQL
Central DB layer
⌁ Temporal orchestrates every workflow — retries, idempotency, recovery.
Stepped into the lead role
Took over team-lead responsibilities when the prior lead departed — design specs, sprint planning, and mentoring a 4-person team, while staying hands-on in development.
Multi-source ingestion at scale
Built a pipeline consuming purchase orders from Mailbox, SFTP, and Azure Blob Storage — files up to 50,000 lines across JSON, CSV, EDI, and Excel, with multiple schema versions per format.
2 min → under 20 sec
Re-engineered the Spring Boot AS400 integration for automated PO creation, cutting order-creation time from ~2 minutes to under 20 seconds — a >85% reduction.
Event-driven receipts
Built an order-receipt pipeline consuming fulfillment events from Azure Event Hubs, persisting to PostgreSQL, and dispatching real-time API notifications and email acknowledgements to partners.
Temporal orchestration
Migrated all long-running workflows to Temporal — ingestion, order creation, receipts, retries, idempotency, and error recovery across the order lifecycle.
Reliability & maintainability
Built a centralized DB service layer (cut duplicated DB code ~25–30%) and led production debugging via structured logging and workflow introspection (reduced manual ops ~20–25%).
Education
M.S. in Computer Science
University of Central Missouri · Aug 2023 — Dec 2024
GPA 3.6/4.0 · ML, AI, Cloud Computing, Advanced Algorithms
Education
B.Tech in Electronics & Communications
Vel Tech University · Aug 2019 — May 2023
GPA 9.14/10 · IEEE Signal Processing Society — President
Certified — AWS Certified Cloud Practitioner
Research
The differentiator
Six published papers in optimization, classification, and quantum-inspired ML. The through-line: squeezing better performance out of learning systems — sometimes with ideas borrowed from quantum computing.
Most-cited · quantum-inspired ML
Quanvolution Neural Network to Recognize Arrhythmia from 2D Scaleogram Features of ECG Signals
A hybrid quantum-inspired CNN that recognizes arrhythmia from ECG scaleograms. My most-cited work, and the clearest example of the quantum-inspired ML angle.
ICITIIT 2022
Research internship
Research Intern — Machine Learning
MITACS · Jul 2022 — Oct 2022 · Kamloops, BC, Canada
- Built hybrid ML models combining quantum-inspired methods with metaheuristic optimization for medical anomaly detection.
- Implemented supervised pipelines in MATLAB, reaching ~95% accuracy on breast-cancer classification.
- Benchmarked 20+ optimization algorithms with statistical validation (F-test, Z-test).
→Proof — MITACS certificate on LinkedIn
6 publications total
- 01
Quanvolution Neural Network to Recognize Arrhythmia from 2D Scaleogram Features of ECG Signals
ICITIIT 202225 cites - 02
Feature Selection Using Elephant Herding Optimization Hybridized with Grey Wolf Optimization for Anomaly Detection in Wafer Manufacturing
Springer — Advanced Computational Paradigms, 2021 - 03
Investigation on Enhancing Binary Classification Accuracy of Supervised Classifiers Using Various Transforms
IOP Conf. Series: Materials Science & Engineering, 2021 - 04
A Study on Real-World Implementation and Future Trends of Internet of Things
IEEE I-SMAC 20206 cites
Skills
The toolbox
Languages
Java·Python·SQL·C·R·MATLAB
Backend & Distributed
Spring Boot·REST APIs·JPA / Hibernate·Temporal Workflows·Microservices·Event-Driven Architecture·Azure Event Hubs·Apache Kafka·PostgreSQL·MySQL
Cloud & DevOps
Docker·Kubernetes·Helm·CI/CD·Jenkins·Git·AWS (S3 · EC2 · Lambda · SageMaker)·Azure (Event Hubs · APIM · Key Vault)
AI / ML
TensorFlow·PyTorch·Scikit-learn·Pandas / NumPy·Deep Learning·LLM API Integration·Prompt Engineering·Model Monitoring & Drift Detection
Off the clock
The person, not the résumé
I think in systems on and off the board. Chess is where I practice the boring fundamentals — patience, calculation, and the humility of a position I misjudged.
Turns out they're the same muscles I use debugging a workflow at 2 a.m. — slow down, read the whole board, and don't fall in love with your first idea.
Off the clock I'm usually learning something new or building one of my own ideas. The long game is to build things that are mine — a job is a chapter, not the whole story.
1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5 — the Italian, on loop
07 — Contact
Let's build something worth shipping.
Open to roles and collaborations across distributed systems and applied AI. The fastest way to reach me is email.