Three products. Designed and directed alone.

A private job-search desk, a research board for India’s National Stock Exchange and a free fitness assessment, each showing its reasoning.

Company: Own products
Role: Product designer and director
Team: Solo
Timeline: Aug 2026 to Oct 2026
Status: Two live, one demo
Outcome: 1,600+ stocks scored daily, no usage data yet
Built with: Python, React, TypeScript. AI: Claude Code as builder

An earlier build of each lab, in under a minute.

Confident answers, with the reasoning stripped out.

Job products, stock screeners and fitness onboarding hand people a confident answer without the reason.

I never ran out of job links. I ran out of hours.

Those hours went to checking which roles were real, whether remote included India, what to say, and who to write to.

job-lab job detail for a senior mobile UX designer role: a next-move strip checking role fit, resume, a person to contact and the send, then whether the posting was written for a real team — In job-lab, those checks sit on every role: fit, resume, a person, the send.

In job-lab, those checks sit on every role: fit, resume, a person, the send.

The missing piece was judgment.

job-lab scores whether a role is worth your hour and ties that score to the next action.

job-lab Today on a phone in the public demo: a banner saying every company, person and number is invented, a morning greeting with 12 things worth doing, counts of 10 roles you can take, 5 live applications and 5 people you know, then Start here with the role to apply to next and the follow-ups due — job-lab’s Today view, in the public demo.

job-lab’s Today view, in the public demo.

Every score has to show its doubt.

market-lab scores the NSE and never lets a score travel alone: risk and liquidity sit beside it in every row.

market-lab Screen on a phone: the session status, a warning that scores use trend and momentum only today, a filter leaving out high-risk stocks, then the first stock with its research score beside its fit, its risk level and its liquidity — market-lab’s Screen view.

market-lab Methodology: what this is and what it is not, saying it is a research tool and not investment advice, a recommendation or a solicitation, and that nobody behind it is registered with SEBI — Its Methodology view says what a score is not: advice.

Its Methodology view says what a score is not: advice.

Every recommendation shows its work.

fit-lab asks only for measurements that change the answer, checks whether your goal and timeline agree, and puts the evidence and limits next to the claim.

fit-lab Result on a phone: where you stand, a waist percentile against the US national health survey with an estimated body-fat range and waist-to-height ratio, then the goal verdict, 8 kg in 12 weeks is a good target, and the reasoning behind it — fit-lab’s Result view.

fit-lab safety stop after chest pain is ticked: Stop exercising for now and get medical clearance, the one place fit-lab will not hand you a plan — Chest pain ticked: stop, see a doctor.

fit-lab safety stop for someone with diabetes who does not exercise yet: Get medical clearance first, then start at light to moderate effort — Diabetes, not yet training: clearance first.

Diabetes, not yet training: clearance first.

Three working products: two live, one public demo.

Each ships with automated tests and accessibility checks; on market-lab, automated WCAG 2.2 A/AA checks found no violations in the views tested.

public job-source families normalized

The real desk is private; the demo uses fictional data.

1,600+

NSE stocks scored every trading day

Research quality, not returns. An independent project, not affiliated with or endorsed by NSE, BSE or NSE Indices.

300+

unit checks across fit-lab’s engines

No usage analytics yet.

Next, I measure real use.

Tests prove the products work; only real use proves they help. Next: reply and interview rates for job-lab, privacy-respecting analytics and clinician testing for fit-lab, guides beyond English, and a log of why each market-lab score moved.

Beyond the brief

There was no brief, so I took on all of it: product framing, scoring models, safety flows, data boundaries, accessibility, tests and deployment. Claude can write a stock screener. It will not decide on its own that every score has to show its doubt. That rule is the product.