Hi, I'm Tharun.
I build AI products
people actually use.

Job seekers spend hours rewriting resumes for each role, and most still fail ATS parsing. Tailoring is repetitive, error-prone, and demoralising.

ATS optimization and per-role customization surfaced in 7 user interviews as the top pain — a solvable, high-frequency job to be done.

Designed and developed the application end-to-end—from user research, PRD, product ideation, prioritization and UX design to frontend architecture, prototyping, state management, and data modeling.

• Role-aware resume tailoring with LLM-based rewriting
• ATS-friendly DOCX export
• Cover letter generation grounded in the JD
• Iterative user feedback loop

RICE + tight MVP scope beat feature breadth. Early users care about export quality more than model choice.

Multi-resume library, JD-level scoring, recruiter-shareable link.

Generic fitness apps push static plans. Users drop off when plans don't reflect real progress, recovery, or preferences.

7 fitness professionals validated that adaptive personalization — not more content — is the differentiator.

Designed and developed the application end-to-end—from user research, PRD, product ideation, prioritization and UX design to frontend architecture, prototyping, state management, and data modeling.

• Goal-aware workout intensity adaptation
• Nutrition recommendations tied to training load
• Closed beta with 25 users for recommendation-quality validation

Closed betas surface adherence patterns you can't get from surveys. Recommendation trust is earned through explanation, not accuracy alone.

Wearable integration, weekly review agent, coach-in-the-loop mode.

Boeing brief: identify high-value LLM applications for commercial pilots within cockpit workflows and constraints.

Aviation is a high-stakes, low-tolerance domain — the right AI use case matters more than the fanciest model.

AI lead on a 5-person MBA consulting team; owned research + application workstream.

• AI opportunity discovery workshops
• 5 solution concepts mapped to business objectives
• Stakeholder-facing recommendation deck (under NDA)

Scope discipline is the deliverable. Problem-first framing beat solution-first every time.

Studying for a certification, learning a new language, and job hunting all pull on the same hours — and each lives in a different app. Nothing shows one clear answer: what matters today, and am I on pace this week?

I wanted to build a single workspace that transforms scattered efforts into a repeatable weekly workflow. The focus was not just on tracking tasks, but on helping users build consistency through planning, reflection, and measurable progress.

Designed and developed the application end-to-end—from user research, PRD, product ideation, prioritization and UX design to frontend architecture, prototyping, state management, and data modeling.

• Weekly sprint-based planning for focused learning
• Daily progress tracking with completion metrics
• Mission Control dashboard for analytics and streaks
• Integrated job application tracker
• Data export for portability and backup

Building NorthStar reinforced the importance of designing around user behavior rather than features. I learned how thoughtful data modeling, reusable component architecture, and meaningful progress metrics can create a product that encourages long-term engagement instead of short-term productivity.

Mobile-first experience Calendar and reminder integrations Goal templates and reusable sprint plans AI-powered weekly insights and recommendations Team accountability and shared learning spaces