Backend Software Engineer, Auth0 (Okta)Ankit

Ankit Phondani

I build the parts of Auth0 that other companies build on.I'm a backend software engineer at Auth0 (Okta), working on Auth0 Teams: tenant lifecycle, multi-tenant data, and the public API that enterprises automate against.

Work line

Auth0 (Okta) since November 2025, then the stops before it, back to 2019.

  1. Auth0 (Okta)

    I work on Auth0 Teams, the product enterprises use to manage all their Auth0 tenants, members and access in one place. My part is tenant lifecycle, multi-tenancy and the public API on top.

    Since November 2025
    1. Teams public API

      The API enterprises use to manage their tenants, members and access as code. I worked on it through to go-live: production readiness planning, the load tests that set its throughput baselines, and the infrastructure plan for launch. Once customers automate against it, keeping its contract stable is the job.

      Public API, go-live
    2. Tenant data

      The multi-tenant data layer under it: cursor pagination with a stable order over datasets taking continuous writes, and the automation that creates and retires tenants.

      PostgreSQL, MongoDB
    3. Change data capture

      Change data capture pipelines that stream transactional identity state into Snowflake, so analysis reads a stream instead of the primary databases.

      Go, Terraform, Snowflake
    4. Self-service SSO

      Lets enterprise customers set up SSO for Auth0 Teams on their own, connecting their identity provider over OIDC or SAML. It bridges the gap across 100+ customer environments and tenants and brings sign-in and access for all of them into one central place in Auth0 Teams.

      OIDC, SAML
    5. Observability

      Tracing, dashboards and alerting across core services, and owning the architectural fix after an incident rather than the patch.

      OpenTelemetry, Datadog
  2. Informatica

    Software engineer on two enterprise products.

    2022 to 2025
    1. AI agent workflow platform

      AI agent engineering: I worked on a platform where multi-agent LLM pipelines are composed from agents, memory and external tools. The work won Informatica's annual InfoStar award.

      InfoStar award
    2. Cloud Data Marketplace

      An enterprise product for operating data workflows. Access control driven by deep asset hierarchies, Elasticsearch search for faster data discovery, a Webpack to Vite migration, and sole ownership of the product in my final year.

      Elasticsearch
  3. Freelance on Upwork

    Freelancing during my MCA: 7+ projects, all rated five stars, and Top Rated status, the top 10% of Upwork freelancers. For an augmented reality app I moved search from PostgreSQL to Elasticsearch, kept the two in sync with pgsync, and built the search APIs on top, with Redis alongside. I also built a classifieds platform with location search and full authentication on Node.js, MongoDB and Firebase.

    2020 to 2021
  4. Internship

    In my final year of BCA I built a travel company's hotel and flight booking platform on my own: supplier API integrations, CCAvenue payments, and a payment link generator the team used to collect payments. Node.js, Express and MongoDB with Redis caching, deployed on AWS.

    2019 to 2020

Writing line

Notes on contracts, pagination and data pipelines.

All writing

  1. Keeping analytics off the primary database

    Change data capture moves the reads that do not belong in the request path onto a stream.

    12 Nov 2026, 2 min
  2. Cursor pagination over data that never stops changing

    Offset pagination skips and repeats rows when writes land between pages. A cursor fixes that, but only if the order underneath it is stable.

    12 Nov 2026, 3 min
  3. Self-service SSO in Auth0 Teams

    What single sign-on is, why companies insist on it, and how a customer can set it up for Auth0 Teams on their own.

    12 Nov 2026, 4 min

Products line

Designed, built and run on my own.

  1. MoveAcross

    Finds software jobs at companies that can actually sponsor a visa. Every morning it checks thousands of official job pages across 18 cities in 12 countries, matches each employer against the official Dutch and UK sponsor registers, and shows the evidence behind every match. Users review their matches, apply on the company's own site, and MoveAcross tracks each application with follow-up reminders. Go and PostgreSQL, with its own background workers built on River, a Postgres-backed job queue for Go, and a Next.js front end.

    Go, PostgreSQL
  2. Revido AI

    Turns a written post into short videos and publishes them to YouTube. A Go worker pool built on goroutines and channels runs the CPU-heavy FFmpeg renders without starving the API, and publishing uses OAuth 2.0 scoped to each user's own channel.

    Go, PostgreSQL, FFmpeg
  3. ScriptMotion

    Generates short tech explainer videos from a structured script: narration with text to speech, captions, animation and composition in Remotion. Renders run through a BullMQ queue with controlled concurrency so they stay stable. It publishes to a live Instagram tech channel.

    Node.js, BullMQ, Remotion
  4. ViteServe

    A VS Code extension that runs a Vite dev server inside legacy Create React App projects. It detects the CRA setup, installs and aligns the dependencies, and starts Vite from the editor. 6,000+ installs on the VS Code Marketplace.

    TypeScript, Node.js
  5. BookHive

    Search over the Goodreads dataset, more than a million books. I imported it into MongoDB and built the REST APIs in Go, with indexes designed so partial-match search stays fast at that size.

    Go, MongoDB