Shaswat Naman
Founder & product builder — research, decisions, shipped
Overview
Social Links
About
I'm a product person who can read the code, write the query, and instrument the funnel. My work starts with a specific user problem — confirmed through research, not assumed — and ends with a shipped decision that I can measure.
The clearest example is Nureli, a mental wellness app I founded for adults who grew up emotionally unseen. Before writing a line of code I interviewed 14 psychologists and clinical specialists to confirm the niche was under-served. Today we have 500+ users (200+ weekly active), a freemium model with a paid tier, and a VIT Business Incubator backing. When retention was weak, I didn't guess: I pulled funnel data in PostHog, found onboarding completion at 35% and day-3 return at 30%, combined it with user feedback, identified two root causes, and shipped two targeted fixes. That's the loop I run.
Before Nureli, I was an AI / Product Strategy Intern at HighScores AI, a US-based edtech startup. I found a data observability gap — the error log captured mistakes but nothing confirmed a student had actually closed the knowledge gap — queried 500+ session records in Python and SQL to validate the pattern, and recommended the Revenge Test feature, which was prioritised on the product roadmap.
I'm also one of ~100 Adobe Student Insiders selected from 19,000+ applicants across India (0.5%), running end-to-end product evaluation of Adobe Student Spaces: mapping friction points by frequency and impact severity, and submitting structured findings to Adobe product and engineering teams across India and the US.
Adobe Student Insider
Adobe Student Insider
Period: Oct 2026—Present
Selected as one of ~100 Student Insiders from 19,000+ applications across India (about 0.5%).
- The kickoff
- The program kicked off at the Adobe Education Summit in Noida, where Adobe flew the cohort in for the kickoff and orientation. We spent the time learning from Adobe's VPs, product, growth, design and community leads, and teams from both the US and India.
- What I'm doing now
- Running end-to-end product evaluation of Adobe Student Spaces (Acrobat) with Adobe teams across India and the US:
- Mapping real student workflow patterns against the product's intended use
- Categorising friction points by frequency and impact severity — not just listing complaints
- Translating field observations into prioritised, actionable product recommendations
- Submitting structured findings to Adobe product and engineering teams across India and the US
Founder of Nureli
Founder · Product & Growth
Period: 2026—Present · On the App Store
A mental wellness app for people who grew up emotionally unseen — from research-confirmed niche to 500+ users.
- The problem
- Childhood emotional neglect is easy to miss because nothing visibly happened. Adults who grew up with it often struggle to name what they feel or to believe their needs matter, and most wellness apps are built for people who already have the words. I confirmed this gap was under-served by interviewing 14 psychologists and clinical specialists before writing a line of code.
- My role
- I own the roadmap, run the user research, instrument the funnel, and make the calls on what ships and what doesn't — on a team of five.
The Nureli team.
- 500+
- users (200+ weekly active)
- 5
- person team
- VIT-TBI
- incubated
Our approach
- Research before roadmap — 14 psychologist interviews confirmed the niche. Every major product decision since has been tied to user feedback or funnel data — not intuition.
- Instrumented from day one — Funnel tracked in PostHog. When retention was weak, I found onboarding completion at 35% and day-3 return at 30%, diagnosed two root causes from data + user interviews, and shipped targeted fixes.
- Gentle, never clinical — The message is 'just see it, nothing to fix.' Reflect, just say it, complain, or reach your inner child — people start wherever they can.
- Safety is a product decision — If someone may be in crisis, Nureli surfaces the 988 Lifeline before anything else. That's a deliberate design and policy call, not a default.
AI / Product Strategy Intern
HighScores AI
May 2025—Jul 2025 · Remote (US)
Edtech startup based in the US. Worked on AI-driven product strategy — analysed learning data, identified gaps, and contributed recommendations to the product roadmap.
highscores.aiHackathon
iQOO Hackathon 2026
Chennai · 12–13 Sep 2026
2nd Runner-Up, Student Track · ₹50,000
Confession: I almost never do hackathons. Forty-eight hours of sitting still? Not my thing. I physically can't.
Then my best friend talked me into one. Somehow I said yes, and that's how my first-ever hackathon happened.
It wasn't a small one. 13,000 people applied nationally, 120 teams were selected, and we were in Chennai for the grand finale.
We built Saathi — a PCOS companion that turns fragmented symptom data into a structured clinical report for the doctor — and came 2nd Runner-Up. The product decision that won us points: we identified the doctor handoff gap as the highest-value problem, not the symptom-logging UX, and built to that.
We built: Saathi, an on-device PCOS companion with a structured clinical handoff report
- 13,000+
- applied nationally
- 120
- teams in the grand finale
- 2nd Runner-Up
- Student Track · ₹50,000
Problems I Work On
Every project here started with a problem I couldn't find a good solution to, and a decision about how to build the right one. These are some of them. For more, GitHub is a good place to start.
Alpha Agents
Financial research is slow because every claim needs to be traced. AlphaAgents sends five specialised agents — Fundamental, Technical, Sentiment, Critic, and Synthesis — running in parallel. A Critic agent adversarially challenges every claim before synthesis, and every material assertion in the final report is cited to a source document, chunk, and relevance score.
Saathi
The problem: PCOS is often missed because symptom data is fragmented across patient, family, and doctor. Saathi lets patients record symptoms privately on-device and generates a structured 3-page clinical report for the doctor handoff — converting longitudinal data into something a clinician can act on in one tap.
Jyotish AI
Most AI answers in devotional and astrology contexts are hallucinated. Jyotish AI answers only from trusted, curated sources — and when it doesn't know, it says so. Guardrails and evals are first-class, not afterthoughts.
Nirnay 112
Emergency operators in India handle Hindi and Hinglish calls where seconds matter. Nirnay 112 identifies the emergency type, key location signals, and recommended response — giving the operator a structured decision, not a transcript.
GitHub Contributions
535 contributions in the last year on GitHub.
