Saransh Ahuja

summary

Zippex was a production service I operated for two and a half years: 5,000+ users at 99.9% uptime on one Firebase backend behind five surfaces, with scheduled sweeps that expired unfilled orders and released the payment hold. Kestrel is a pipeline designed for a 16 GB machine: chunked ingest into a staging catalog promoted only after the day verifies, deterministic replay proven by a streaming digest, and a VM plus systemd plus Tailscale deployment written up but not provisioned. Daneel runs on scheduled Cloud Functions, enforces kill rules in code, and writes its ledger exactly once. IEEE-published at CAI 2023, Bachelor of Computer Information Systems from UFV.

core competencies

• Backend Operations & Uptime
• Data Pipelines & Streaming Ingest
• Deployment Design (VM, systemd, Docker)
• Firebase, Cloud Functions & Firestore
• Scheduled Jobs & Idempotent Sweeps
• Secrets & Access Boundaries
• Deterministic Replay & Verification
• Run Records & Post-Mortem Reporting

experience

• Zippex Technologies Inc

  Developer and Operations Project Manager · Abbotsford, BC

  Feb 2024 - Aug 2026
  • Built and launched a delivery platform using React Native, scaling to 5,000+ active users with 99.9% uptime
  • Designed scalable backend with Firebase and Stripe, processing 1,000+ transactions/month with zero downtime
  • Operated one Firebase backend (Cloud Functions, Firestore rules, Stripe webhooks) shared by the customer app, driver app, merchant, admin, and marketing sites
  • Built scheduled Firestore sweeps that expire unfilled orders at 20 minutes, release the Stripe hold in-process, and flip silent drivers offline after 45 minutes
  • Integrated Google Maps API for real-time tracking, improving delivery accuracy by 30% and cutting average delivery times by 20%
  • Implemented AI routing for delivery optimization, cutting carbon emissions 10% per delivery
• Paragon PMO

  Software Developer · Abbotsford, BC

  Jan 2026 - Jun 2026
  • Built backend APIs and database integrations supporting application workflows and data management
  • Owned the full lifecycle (planning, development, testing, deployment, and post-launch maintenance), improving application reliability
  • Developed full-stack internal PMO tools and client-facing web applications using React and Next.js
• Westland Insurance, IT Integrations (M&A)

  Project Support Analyst · Vancouver, BC

  Nov 2022 - Oct 2023
  • Worked within a pan-Canada team managing network upgrades, hardware deployments, and data migrations for acquired branches
  • Built the integration tracking dashboard for M&A onboarding of 400+ branches across Canada, still in production today
  • Ensured consistency in technical methodology across multiple integration vendors
  • Developed technical documentation and knowledge bases using IT Glue and ServiceNow
• University of the Fraser Valley

  AI Research Lead (AI Researcher May 2022 - Dec 2022) · Abbotsford, BC

  Jan 2023 - Dec 2023
  • Led research on speech emotion recognition using deep neural networks and support vector machines
  • Evaluated model performance using precision, recall, and F1-score metrics
  • Published findings at IEEE CAI 2023
• University of the Fraser Valley

  Junior Project Manager · Abbotsford, BC

  May 2024 - Dec 2025
  • Implemented ACADEM, UFV's international accreditation system, aligning programs with global standards (AACSB, EFMD)
  • Built and managed project dashboards and timelines, increasing on-time milestone completion by 25%

relevant projects

• 
  KestrelA market-making pipeline built to stream a 44 GB day through a 16 GB machine.
• 
  DaneelScheduled paper trading with kill rules the code enforces, reviewed by AI after the close.
• 
  RelistEvery car on the lot, on Marketplace, never stale.
• 

  ZIPPEX

  You need something moved, Zippex moves it.

  Visit Zippex
  ZippexUber Eats for Facebook Marketplace.
• 
  MnemeThe brain for your AI.

• Kestrel: A market-making pipeline built to stream a 44 GB day through a 16 GB machine.
• Daneel: Scheduled paper trading with kill rules the code enforces, reviewed by AI after the close. (daneel.saranshahuja.com)
• Relist: Every car on the lot, on Marketplace, never stale. (relist-ten.vercel.app)
• Zippex: Uber Eats for Facebook Marketplace. (zippex.app)
• Mneme: The brain for your AI. (trymneme.com)