CS student building the future of scouting

Computer Vision and AI Systems engineer passionate about sports analytics, agentic AI, and EdTech. Founder of Mentrixa (AI tutoring SaaS). I help teams understand AI + Computer Vision in sports and build production-scale systems.

• Building agentic AI systems with Claude, Gemini, and Groq
• Real-time sports analytics & computer vision pipelines
• AI automation with Make.com, n8n, and LangChain
• Gamified EdTech SaaS with adaptive learning AI

• Vision Models: OpenCV, MediaPipe, YOLOv8, CNNs, ResNet, Vision-Language Models (VLMs), Object Detection, Image Segmentation
• ML/AI Frameworks: PyTorch, TensorFlow, scikit-learn, HuggingFace Transformers, MLflow, Weights & Biases
• Agentic AI: LangChain, LangGraph, RAG, multi-agent orchestration, Claude API, Gemini, Groq Llama, DeepSeek AI

Horizontal/vertical scaling, load balancing, CAP theorem, CQRS, cache invalidation, message queues, microservices, multi-region failover, rate limiting

LangChain, LangGraph, RAG, multi-agent orchestration, Make.com, n8n, prompt engineering, reproducible experimentation

Python (FastAPI), Node.js, Docker, Kubernetes, AWS (EC2, S3, Lambda, SQS), Supabase (Postgres, Auth, Realtime), Redis, PostgreSQL, MongoDB, Snowflake, NATS/JetStream, RabbitMQ

Next.js App Router, React, Tailwind CSS, TypeScript, Vercel, Stripe, Resend, PostHog, Sentry APM, WebSocket, REST APIs

Python, TypeScript, Go, C++, Java, SQL, Bash

Docker, Kubernetes, Helm, Jenkins, GitHub Actions, Git, Linux, pytest, Playwright, Twilio, ElevenLabs

Gamified EdTech SaaS: AI tutoring marketplace, adaptive practice engine, competitive ranking system

• Shipped full-stack SaaS solo in 8 months while enrolled full-time in second-year CS
• Architected horizontally scalable multi-tenant system: Next.js 16, Python FastAPI, Supabase Postgres (87 migrations + Row-Level Security), Drizzle ORM, Stripe Connect, Resend, AWS S3, Sentry APM
• Integrated Claude API + Google Gemini 2.5 for adaptive AI quest generation and personalized practice engine
• Achieved 14/14 Playwright E2E tests + 44/44 Vitest unit tests passing; deployed to Vercel with 11 cron jobs, edge auth, Turnstile CAPTCHA, rate limiting
• Designed real-time skill duels, XP leveling, division leaderboards using Supabase Realtime for sub-second score propagation and Postgres FTS for fast search

Autonomous software platform: builds, deploys, monitors, and self-heals production apps from plain-English intent

• Architected end-to-end autonomous factory: intent classification → architecture planning → LLM code generation → Docker containerization → horizontal deployment → real-time monitoring
• Engineered self-healing closed-loop repair system: monitors deployments, diagnoses via LLM reasoning, generates fixes, rebuilds Docker images, redeploys autonomously
• Applying circuit-breaker patterns, multi-region failover logic, immutable audit trails for every agent action

Upwork | 7 production clients across education, recruiting, sports-adjacent SaaS

• Delivered 75% automation success rate uplift, 40% client ROI within 90 days
• Engineered stateless Python FastAPI microservices with RabbitMQ async queues, horizontal scaling, circuit-breaker fault tolerance on AWS EC2
• Increased qualified lead conversion by 35% with Python RAG workflows (LangChain, ChromaDB) with TTL-cached embeddings, write-through cache invalidation
• Reduced manual B2B workflow intervention by 45% with LLM-based classification, output guardrails, exponential-backoff retry logic, Helm-packaged microservices across 7 client environments

Montreal, QC | Sports analytics startup building real-time computer vision for professional basketball

• Achieved 92% basketball-event detection accuracy at 30 FPS with OpenCV + MediaPipe (<100ms CPU inference latency) on AWS EC2 streams (200+ matches/season)
• Reduced analyst review time by 30% per match with stateless Python FastAPI REST services + WebSocket streaming to React dashboard (8 analysts)
• Sustained 99.9% pipeline uptime across full professional season with staged ingestion-processing-inference backend, automated Bash deployment scripts, zero-downtime updates

FastAPI, PostgreSQL, Redis, Groq Llama 3.1, Claude, Next.js

• Sub-200ms response times under concurrent agentic workloads via NEXUS engine (parse intent → route to specialized agents → write-through Redis caching)
• Immutable audit trail to PostgreSQL via Alembic for every agent action with human-in-the-loop budget enforcement

FastAPI, React, MongoDB Atlas, Snowflake, Google Gemini, OpenRouter, WebSocket

• Real-time incident triage from text/images with Google Gemini multimodal reasoning + Snowflake Cortex predictive forecasting
• OpenRouter multi-model routing with automatic failover across providers based on latency/cost/reasoning depth

Python, FastAPI, NATS/JetStream, Docker, Kubernetes, Helm

• Won Best Use of Snowflake API, Invest Ottawa Hackathon 2026
• 99.99% message delivery reliability with NATS/JetStream pub-sub in Docker/Kubernetes, idempotent message processing

FastAPI, Streamlit, DeepSeek AI, Plotly 3D, Twilio

• Real-time demand prediction engine with write-optimized FastAPI ingestion and constraint-aware scheduling
• Interactive Streamlit dashboard with Plotly 3D visualizations + Twilio SMS alerting

• AWS Fundamentals of Machine Learning and Artificial Intelligence Amazon Web Services | Oct 2025
• Intro to Machine Learning, scikit-learn, Random Forest Kaggle | Nov 2025
• Practical GitHub Actions LinkedIn Learning | Feb 2026
• Introduction to Artificial Intelligence LinkedIn Learning | 2025
• Docker and DevOps Fundamentals LinkedIn Learning | 2026
• AWS Cloud Practitioner Amazon Web Services | In progress (2026)

• Best Use of Snowflake API Invest Ottawa Hackathon 2026 (Chronos Cloud distributed digital twin platform)
• Faculty of Engineering Achievement Award University of Ottawa 2024-2025
• 68 NeetCode DSA Submissions Arrays, graphs, dynamic programming, trees, system design patterns
• GitHub Pull Shark & YOLO Badges Consistent open-source commit activity and PR contributions