Product Manager & Hands-on Builder shaping AI learning products across a two-sided school platform and a student app. Owns roadmap delivery across teacher co-pilot workflows, AI tutoring, and criterion-based grading, while pair-building with engineering on model integration and retrieval. Drove onboarding and activation gains by using funnel data to simplify new-user setup and reduce drop-off.
Overview
10
10
years of professional experience
Work History
Product Manager & Hands-on Builder
Koncepts (IB AI Co-Pilot), PrepAiro
01.2026 - Current
Two-sided AI platform for IB schools: a teacher co-pilot (assessment, planning, grading, analytics) and a student app (AI tutoring, practice, gamified learning) across PYP, MYP and DP and ~9 subjects.
Sole product manager for the platform, owning the roadmap end to end across both the teacher co-pilot and the student app, spanning three IB programmes and ~9 subjects, shipping with a 12-person engineering team
Managed platform with 4,000 daily active users, 16,000 monthly active users (25% DAU/MAU) across 89 countries and 600+ educators
Redesigned new-user onboarding flow based on production-funnel analysis, increasing same-day completion from 35% to 53%, reducing pre-grade drop-off from 54% to 35%, and doubling daily active users
Pair-built and shipped the AI tutoring engine with engineering: a RAG-grounded chat tutor, a real-time AI voice tutor, and a photoreal avatar Class Tutor (Claude/GPT/Gemini, Milvus, LiveKit, Anam)
Own the AI Assessment Studio, from IB criterion segmentation generation (Criteria A to D, strands, holistic banding) through auto-grading with evidence-linked feedback, supporting 28,000+ AI-graded submissions across 600+ educators
Led LLM evaluation for grading quality, using teacher corrections as ground truth to inform rubric and prompt changes, ensuring responsible AI; enabled auditing and overriding of scores by teachers
Make the model-selection tradeoffs across Claude, GPT and Gemini, choosing per workload on cost, latency and instruction-following
Own the broader student learning surface (notes, quizzes, gamified practice), drive retention from engagement data, and drove the 0-to-1 PYP programme launch to parity with MYP/DP plus the slide-deck builder with live in-class quizzes
PM
ClassPlus ( Polaris School Of Technology )
05.2025 - 12.2025
0-1 Placements Outreach Design
Led on-the-job training for 12 batches (~350 learners), enrolling 8 hiring partners and achieving 80% placement of learners into tech roles at median ₹ 6 LPA package
Co-developed applied-technology curriculum to align with industry tools and demands
Facilitated 8 fireside chats with industry leaders and conducted B2B workshops with influencers, managing partner relationships
Owned product and growth end to end, growing the app to 200K+ organic installs and 500+ paying subscribers while owning scope, roadmap and the acquisition-to-conversion funnel
Delivered core AI learning features, including Airo AI mentor (chat and real-time voice), AI answer evaluation for descriptive answers, adaptive learning paths, mock tests, and practice quizzes, enhancing user learning experience
Drove engagement and retention with gamified features (duels, daily challenges, streaks, leaderboards) and in-app referral loops that fueled organic growth
Collaborated with engineering to develop core AI features: Airo voice mentor (LiveKit), AI answer evaluation, and RAG retrieval, while conducting discovery and A/B experimentation (Firebase Analytics, Remote Config) to optimize product performance
All-in-one UPSC exam-prep app with a personal AI mentor (Airo).
Founder
history0and1 (UPSC prep platform)
01.2023 - 12.2023
Developed a tool for structured feedback on descriptive answers for UPSC mains.
The mains evaluator is the direct ancestor of everything since: criterion-based scoring of long-form written answers, the same problem I now solve at IB scale in Koncepts and benchmark publicly in GradeBench
Ran it solo end to end, product through content, and sourced the problem from my own failure, which is still how I choose what to build
Managed all aspects of product development, from content creation to user experience, directly addressing challenges identified from personal experience.
Transitioned from solo project to collaborative environment at PrepAiro, continuing to tackle the same evaluation problem with a team focus.
Education Content Creator & UPSC Candidate
06.2020 - 12.2022
Spent two and a half years in exam-prep market, building user-focused products as a former user turned PM.
Created educational audio series for Kuku FM, a paid consumer learning product with a national audience
Two and a half years inside the exam-prep market I went on to build products for. I was the user before I was the PM.
Cleared SSC CGL (Tier 3), SSC CHSL and LIC-HFL, and UPSC prelims, from applicant pools in the millions
Identified key challenges in descriptive answers for candidates at history0and1, Airo, and Koncepts, addressing the lack of useful feedback at scale.
Software Engineer
QuantLattice Systems
08.2016 - 05.2020
Developed and maintained Java / Spring Boot backend services and REST APIs, enhancing platform functionality
Created Python ETL pipelines to ingest, validate, and transform data from SQL, REST APIs, and flat-file sources, ensuring data integrity
Designed and optimized PostgreSQL schemas, using indexing and partitioning to reduce query latency
Containerized and deployed services on AWS (EC2, S3, RDS) using Docker and CI/CD, streamlining deployment processes
Education
B.Tech. - Computer Science
United College of Engineering & Research (UCER)
Prayagraj
01-2016
Skills
LLM integration (Claude, GPT, Gemini)
RAG pipelines and vector search (Milvus)
Prompt engineering
Real-time voice agents (LiveKit)
TTS (ElevenLabs)
Product analytics
A/B testing and experimentation (Statsig, GrowthBook, Firebase Remote Config)
Onboarding funnel optimization
Retention analysis
Cohort analysis
Roadmap management
Cross-functional delivery
Prioritization frameworks
B2B partnerships
Two-sided platform strategy
Accomplishments
Sole PM on a platform at 4,000 DAU / 16,000 MAU across 89 countries - 25% stickiness, three IB programmes, ~9 subjects, 600+ educators, one 12-person engineering team
28,000+ submissions AI-graded against IB criteria, with evidence-linked feedback teachers can audit and override
200K+ organic installs, 500+ paying users - built with content, referrals and social loops, zero paid spend
Engineer who became a PM and kept building - 4 years Java/Python before product; writes the prompts, reads the code, argues tradeoffs with engineering