Dev Sapariya

Hi, I'm Dev 👋🏻

I'm a software engineer who likes turning ideas into working products, especially when I get to learn a new technology along the way.

Right now, I work at Sage AI Labs on data infrastructure and the internal tools our retail operations team uses every day. I graduated in 2025, but a lot of what I know I learned outside the classroom: building projects, breaking things, and figuring out how to make them work in the real world.

Read how I got here →

Work experience
See all work
  • Jan 2025 – Present
    Sage AI Labs
    Software Engineer · Remote
    • Onboarded 55+ e-commerce retailers end to end, from writing the product data extraction scripts to validating the data and plugging each source into pipelines that process more than 100K products a day.
    • Building an AI harness that writes those extraction scripts on its own. A script that takes 1–72 hours by hand now takes 30 minutes to 2 hours, it can work on 4 retailers in parallel, and 165+ of its scripts have been merged into the data pipeline so far. Each one is checked against the live store and goes through a human approval step before it ships.
    • Built serverless image-processing pipelines on Modal and Cloudflare R2 that handle over 1 million images a day. Replacing per-image R2 existence checks with a database cache cut R2 Class A operations by about half and made runs faster and lighter on compute.
    • Designed and built a database-driven scheduler that replaced 200+ separate Modal cron jobs, so the operations team can change scrape schedules from the ops console without waiting on a deploy.
    • Built a live-inventory service that checks current size availability and price for 58 non-Shopify retailers at checkout, plus a shipping-cost pipeline and API that now run in production.
    Read more
  • Aug 2024 – Nov 2024
    Styl · Acquired by Sage AI Labs
    Backend Developer, Part-Time · Remote
    • Joined Styl part-time while I was still in college and built a one-click scraping pipeline on a VM that ran the team's 40+ retailer scrapers (Python and Node.js). It replaced running scrapers locally and uploading CSVs by hand: a MongoDB-backed scheduler refreshed each brand every 72 hours, then cleaned, deduplicated, and loaded the results.
    • Wrote validation rules for the scraped product data, covering duplicates, size taxonomy, price bounds, image counts, and review cleanup, since a third-party ML service read the catalog directly from our MongoDB.
    • Built a Flask admin dashboard for reviewing each brand's data and moving it into the live catalog. It kept existing item IDs, flagged changed prices, sizes, and reviews, soft-deleted removed products, and used OpenAI to shorten product descriptions.
    • Containerized the pipeline and an image-compression worker with Docker and deployed them to the VM through GitHub Actions, waiting for in-progress database writes to finish before replacing the container. I also built the Flask API for a brand analytics dashboard, with JWT authentication, per-company access control, response caching, and MongoDB aggregations of user swipe data.
    Read more
Recent projects
See all projects
Let's connect

If you want to talk about something or just say hi, reach out on social media or send me an email.