Anurag Jha
Black and white portrait of Anurag Jha in a black t-shirt against a dark background.

AI Product Manager · Director of Product Management

Anurag Jha

I build AI products that move the numbers.

At NeuroDiscovery AI I lead six AI products and a team of 13. Over six years I have shipped 20 products, 16 of them AI first, across health, fintech, automotive, education and gaming. I take each one from a prototype I build myself to production, and I measure what it changes.

products shipped, 16 AI firstProduct index
20
loans through the journey I rebuiltCompany figure, lifetime
$400M+
installs of the AI app I ledCompany figure, lifetime
3.5M+
clinical records through my LLM platformProduction database
2.5M

Impact · 2019 to 2026

The numbers, with their sources.

20products shipped, 14 as owner or founderProduct index, 2019 to 2026
$400M+education loans disbursed on the Yocket platform whose loan journey I rebuiltCompany figure, lifetime (INR 3,500+ crore)
3.5M+installs of the Jivi consumer AI app I ledCompany figure, lifetime; 500K+ on Google Play during my time
~$2.4Mof Spyne's ARR came from the AI imaging line I owned, about 80% of ~$3MCEO interview, Mar 2023; my arithmetic
$1B+car transactions at OLX Autos, where I helped launch AI self inspectionCompany press, Aug 2021
~3.5Mpeople reached across the products I worked onEstimate: Jivi installs plus Yocket, AIcade and First Draft users
2.5Mclinical records processed by NeuroAudit, with 23K+ reviewer decisionsProduction database, to Jul 2026
~8 wksmedian from start to launch across my four dated launches10 to 131 days: NeuroAudit, Clinical Trials Portal, Neura, ProdMan.AI
5industries: health, fintech, automotive, education, gamingEmployment and public records

Verified recordCompany figureMy figureEstimate

Products, flagship first.

All 20 products
Jarvis, the assistant I run my work on. An AI generated frame from a video I made about it.

Claude Code · agents · evals

The system I built to run my week.

I run my own product work through Claude Code agents: meeting notes become tickets, tickets become prototypes inside the real codebase, and nothing ships until it passes its evals and my review.

  1. Stage 1CaptureNotes in, journey tickets out
    1. 01
      Intake

      Meeting notes, chat dumps and Jira items are logged word for word and sorted by product and type.

      Gate: Low confidence asks one question
    2. 02
      Context

      The agent reads the product map, the entry points and every page a change touches, before writing anything.

    3. 03
      Ticket

      Written as a user journey: where the user starts, what they click, what they see, the API call, the response, where they go next.

      Gate: Ticket eval: testable criteria, zero personal data
  2. Stage 2BuildPrototypes inside the real product
    1. 04
      Design system

      Prototypes reuse the real product's components. No new patterns sneak in.

    2. 05
      Prototype

      Built inside the living product repository, on its own branch.

      Gate: Build, types and the full test suite must pass
    3. 06
      Approval

      I review side by side. Every correction is written back into the skill so it never repeats.

      Gate: Nothing reaches Jira without my go
  3. Stage 3ShipSame repository, tested
    1. 07
      Handoff

      Locked scope, acceptance criteria, edge cases and test hooks for the engineer's branch.

    2. 08
      Release

      Engineers build on the same repository, so the prototype and the product share code.

    3. 09
      QA and evals

      Every acceptance criterion maps to a test row with a named owner and environment.

      Gate: Rows stay blank until a human runs them
  4. Stage 4ShowDemos, notes, review
    1. 10
      Demo

      Interactive product tours in Supademo, for example a twenty step tour of the patient app.

    2. 11
      Release notes

      Drafts shaped for the team, partners or leadership.

      Gate: A draft is not a send
    3. 12
      Review

      A run log records intake, route, output, gate and outcome for every item, and I review it.

0/4 → 4/4ticket eval, from an ungated draft that leaked 5 personal identifiers to a gated draft with noneVerified record · Synthetic test note, not production traffic
5 / 5requests routed to the right workflow on a small golden set of 5, after fixing the one miss in the first runVerified record · Smoke test, not a full eval set
304 / 304of the product's own automated tests still passing after a prototype was built inside the real appVerified record · Run log, 2026
2 wks → <1 wkfrom idea to release, since prototypes moved into the real codebaseMy figure · My own figure
60% → 20%of my week on documentation and chores, a rough estimate rather than a measurement; the rest now goes to strategyMy estimate · Self estimate of time allocation, not a measurement

Side quests, shipped.

Career · 2019 to now

Five companies, one habit: ship and measure.

  1. Jun 2025 to nowEnterprise and partner AI

    NeuroDiscovery AI

    Director of Product Management

    I own strategy and the roadmap for six AI products and lead 2 product managers, 3 designers and 8 engineers. I build prototypes myself in Claude Code, so designers and engineers start from working software.

    2.5Mrecords through NeuroAudit
    5products shipped in 2026
    Tech30YourStory 2025, top 30 of 2,000+ startups

    Products: NeuroAudit · Neura · Reinforcement learning module · Clinical Trials Portal · NeuroTerminal · PM delivery system on Claude Code

  2. Apr 2024 to May 2025Consumer app and B2B partnerships

    Jivi AI

    Senior Product Manager

    I led 2 product managers, 6 developers and 2 designers across the consumer AI health app, hospital partnerships and internal tools, and shipped Dr Jivi, a copilot for doctors.

    17 → 7 minper consultation with Dr Jivi
    3.5M+app installs, lifetime
    #1Open Medical LLM Leaderboard, May 2024

    Products: Dr Jivi · Jivi consumer AI health app · Medx model and medical vision · Multi agent workflow system

  3. Oct 2022 to Apr 2024Marketplace: students and partner lenders

    Yocket

    Product Manager, Product and AI

    I turned an advisor driven education loan process into a self serve journey and shipped Loan GPT for advisors, working with 3 developers, 4 operations staff and 50+ sales colleagues.

    3.1xloan conversion
    7 d → 48 happroval time
    $400M+loans on the platform, lifetime

    Products: Self serve education loans · Loan GPT · Lender and advisor CRM

  4. Aug 2021 to Oct 2022B2B and B2B2C AI imaging

    Spyne AI

    Associate Product Manager

    I owned the automotive AI imaging line, Spyne's biggest vertical, sold as APIs to dealers and marketplaces in five regions.

    ~80%of company revenue from my line
    ~10xARR growth year to ~$3M
    40% → 90%model accuracy

    Products: AI imaging APIs for cars · 360 degree car spin

  5. Feb 2021 to Jul 2021Consumer car selling

    OLX Autos

    Product Intern

    I worked on AI self inspection, where sellers inspect their own car with an AR guided camera, live in India and Argentina.

    45 → <10 minper inspection
    $1B+OLX Autos transactions, Aug 2021

    Products: AI self inspection

Education

  • SP Jain School of Global ManagementMBA, Marketing · GPA 3.2 / 4 · President of the placement committee for 97 students2021
  • KC CollegeBCom and IT · First Division2019
  • DAV Public SchoolCommerce and Maths, 94.6% · School Captain2016

LinkedIn recommendations

In their words.

“As his manager and lead engineer at Spyne.ai, I was inspired by his unwavering vision for our project, driving us to achieve a remarkable 10X ARR growth in just one year. Anurag's strong bias for action and courage to make tough decisions elevated our product's success.”
Ajay MudhaiChief AI Officer, Irame · ex AI Lead, SpyneManaged me directly at Spyne
“Anurag's remarkable work on the organization's biggest vertical, despite challenges, set him apart. From a non-technical background, he mastered technical acumen, scaling products and laying foundations for other verticals. Highly recommended for any AI-related role or project!”
Abhinav SharmaCTO, Intract and Authena · ex SpyneSenior colleague at Spyne
“Anurag is a true catalyst for opportunity, not only identifying and seizing them but also ensuring others benefit. In the AI domain, his knowledge is unmatched, a problem-solver with an array of tools at his disposal.”
Arpit Singh AhluwaliaVisual Content Creator, JCDecaux Middle EastWorked on the same team at Spyne
All recommendations on LinkedIn

Speaking and life

Off the roadmap.

I was a competitive speaker before I was a product manager. It still shapes how I work: say the one thing clearly, then show the proof.

Black and white portrait of Anurag Jha.
  1. Unplugged by TEDxGateway

    Speaker at India's largest open mic for ideas

    Speaker badge.
    On the Unplugged stage.
  2. TEDx at SP Jain, Sydney

    Talk on the Sydney campus, 2022

  3. A.D. Shroff Memorial Elocution, first prize

    Forum of Free Enterprise, for KC College, 2018: "Role of youth in building a new India"

    At the Forum of Free Enterprise.
  4. President, SP Jain placement committee

    Elected in Sydney, 2021; ran placements for 97 students

    Election speech, Sydney.
    Presenting at SP Jain.

Moments

Contact

Let’s build what comes next.

I lead AI product from strategy to shipped software: the roadmap, the prototype, the evals and the numbers that prove it worked. I am open to product leadership roles in AI, and to advising teams putting LLMs into high stakes workflows.