VINAY VALSON

Ideas are everywhere. Execution is everything.

This is the room view. Switch to the page view

Work

  1. Cow Lameness Detection (CV / MLOps)

    Crowdsourced pairwise scoring app plus GPU inference pipeline on OpenStack: Next.js, FastAPI, NATS, YOLOv8, ByteTrack, and DINOv3 for automated lameness screening, with containerized inference and a live demo.

    Open Cow Lameness Detection (CV / MLOps)
  2. Predictive Analytics of External Factors on Clarius Scanner Sales

    Graduate capstone: performed forecasting and exploratory analysis to evaluate how external factors influenced Clarius Mobile Health monthly scanner sales in Canada and US.

    Open Predictive Analytics of External Factors on Clarius Scanner Sales
  3. AI Agent Content Operations

    Python application with content scoring, caption generation, human approval gates, and Instagram Graph API integration.

    Open AI Agent Content Operations
  4. RAG Music Search Platform

    Hybrid retrieval over Amazon Digital Music data combining BM25 keyword search with FAISS semantic search for natural-language relevance improvements.

    Open RAG Music Search Platform
  5. CocoaBoard

    Interactive analytics dashboard with filtering, aggregation, and LLM chatbot-assisted exploration across country, product, and time dimensions.

    Open CocoaBoard
  6. UBC Reddit Reporter

    AWS Lambda serverless app with Amazon Bedrock sentiment analysis; automated UBC community reporting.

    Open UBC Reddit Reporter
  7. Imitation Game

    Python encryption utility implementing symmetric and asymmetric cryptography with shared-secret and public-key interfaces.

    Open Imitation Game
  8. Fake Real News Detection

    Naive Bayes fake-news classifier with NLP preprocessing and feature extraction.

    Open Fake Real News Detection
  9. Customer Reviews Polarity Analysis

    ML sentiment analysis (NLP) for positive/negative product-review classification.

    Open Customer Reviews Polarity Analysis
  10. Developer Portfolio

    This site: Next.js portfolio with Tailwind CSS and a Stitch-inspired UI.

    Open Developer Portfolio
  11. ARgorithm

    ARgorithm is a project aimed at providing educators and enthusiasts a tool to create AR visualizations that explain the workings of data structures and algorithms with ease.

    Open ARgorithm
  12. Patatap Clone

    This is an app created using Paperjs and Howlerjs to make a clone version of Tapatap.com.. My own effort to create one. Will add more colors and more shapes and sounds in the future.

    Open Patatap Clone

Journey

  1. Solutions Engineer, Tailscale

    Experience · September 2026 - Present · Vancouver, Canada

    Helping organizations replace clunky legacy VPNs with seamless, WireGuard®-based zero-trust connectivity. Hands-on architecture spanning multi-cloud environments (AWS/GCP/Azure), IaC (Terraform), and IAM/SSO integrations. Translating customer feedback directly into product requirements for Tailscale's core platform.

    Stack: Tailscale, WireGuard, Zero Trust, AWS, GCP, Azure, Terraform, IAM, SSO

    Tailscale
  2. Graduate Research and Teaching Assistant, University of British Columbia

    Experience · January 2026 - August 2026 · Vancouver, Canada

    Building computer vision and MLOps components for a cow lameness detection pipeline for segmentation of workflows, evaluation metrics, backend-frontend, infra setup, with containerized inference services on OpenStack for scalable workloads. Designed and deployed the full-stack platform on OpenStack Arbutus: Next.js + FastAPI + Postgres + NATS job queue to a vGPU worker (YOLOv8, ByteTrack, DINOv3), with a live demo at https://cow-app.tail1c5b2c.ts.net/. TA for Cloud Computing: ran tutorials and office hours for 60+ graduate students on distributed systems and deployment topics, and coached 12 capstone teams through architecture and design reviews toward more production-ready cloud deliverables.

    Stack: Python, Next.js, FastAPI, PostgreSQL, NATS, YOLOv8, ByteTrack, DINOv3, Computer Vision, MLOps, Docker, OpenStack, CI/CD, LLM workflows, RAG, Agent orchestration, GitHub

    University of British Columbia
  3. Master of Data Science, University of British Columbia

    Education · 2025 - 2026

    Academic Representative | Scholarship | ML, LLM, AI, Python, R, Statistics

    Cumulative Average: 91.2

    University of British Columbia
  4. Software Engineer (Remote), Blue Rabbit Group Ltd.

    Experience · April 2024 - November 2025 · Hong Kong / India

    Joined as the first software engineer and shipped the Series A MVP end to end. Built data pipelines, real-time analytics workflows, and AI-assisted automation for portfolio and trading operations, reducing manual risk review workload across investment workflows. Shipped the product using Django and React on GCP for 50+ clients, including order placement and client dashboards with sub-second median response times for positions, risk, and market views; integrated DBS Bank APIs. Designed AI agents with human review gates that automated analyst research and asset risk scoring, cutting analyst workflow time by roughly 70% without added headcount. Drove the migration from monolith to microservices (Next.js, NeonDB, Inngest, Clerk on Railway), automated monthly personalized reports for 50+ clients, instrumented product funnels and A/B experiments with PostHog, and orchestrated agent-style analytics/reporting workflows with validation steps before operational use.

    Stack: Python, Django, React, Next.js, PostgreSQL, NeonDB, Microservices, REST APIs, GCP, PostHog, LLM, Slack Bots, Inngest, Clerk auth, Railway

    Blue Rabbit Group Ltd.
  5. Staff Consultant | Software Engineer, Oracle Financial Services Software Limited

    Experience · August 2021 - April 2024 · Bangalore, India / Hong Kong

    Led onsite client-facing delivery for Oracle Banking Digital Experience (OBDX) in Hong Kong across fraud detection, FX dashboards, payments, QR gateway, and open banking; promoted to Staff Consultant within a year. Built multi-tenant open banking and access management capabilities that reduced partner onboarding effort by approximately 40% compared with the prior process. Translated customer requirements into workable PRDs and production code; gated changes on POCs plus re-signed PRDs rather than closing tickets as working as designed. Received Rising Star and Ever Ready awards; mentored and onboarded software engineers to the team.

    Stack: Java, Eclipse, OJET, JPA-ORM, SQL, Oracle DB, SOAP/XML, OBDX, Apache FOP

    Oracle Financial Services Software Limited
  6. Bachelor of Computer Engineering, Savitribai Phule Pune University, Army Institute of Technology

    Education · 2017 - 2021

    Scholarship | First Class with Distinction

    CGPA: 8.91

    Savitribai Phule Pune University, Army Institute of Technology

Info

Our family got its first Intel desktop in 2005, and that was pretty much it for me. I spent years trying to beat games the hard way, hunting edge cases, shaving seconds off puzzles, and poking at anything that felt like a secret. Eventually the fun shifted from clearing levels to figuring out how the machine was pulling it off. Taking PCs apart and putting them back together became a habit. I still default to that same move when something breaks in life. I strip it down, find the real bottleneck, then build up again. Engineering school just gave that instinct a vocabulary.

Outside of work I play drums, love to travel, chase photos wherever I end up, and keep a second brain so I do not lose the stories. The gallery on this site is mostly Banff and the Rockies. Rhythm, light, and systems all feel like the same itch to me. I notice the pattern, then make something with it.

Based in Vancouver, Canada.