Hi, my name is Rakshit

and I love building

My Story

LearnApp · 2018–2020 · Core Team Member

I joined LearnApp because as a college student I believed that financial education was broken in India and I believed that the only way to bring change in financial habits was through education.

  • Designed library of investing and finance courses
  • 50% course completion rate (category average: ~10%)
  • ₹1.2Cr ARR
Ask my agent about the LearnApp's course's module →

INDMoney · 2020–2022 · Associate Product Manager

Joined INDMoney because I realised that while financial education was important, it only changes what people know. But the real leverage in personal finance is at the moment of decision when you're buying a product, not before it.

  • Built financial planning module with recommendation algorithms — 4–5 hours to 15 minutes per plan
  • Enabled 20,000+ advisory plans
  • Built personalized insurance recommendation engine aligned to user goals — ₹1.5Cr ARR in year one
Ask my agent about the financial planning module at INDMoney →

ET Money · 2022–Present · Product Manager

Four years at ET Money taught me one thing about AI products: the model is interchangeable, the eval is the product. The rubric is where the judgment lives. It's also where I now spend most of my time.

  • Built AI support automation — 17K to 7K monthly tickets via intent classification, 100+ resolution workflows, and a four-dimension eval pipeline (empathy, accuracy, policy, regression)
  • Built 0-to-1 offline distribution business for relationship managers — 12% of SIP book, 10% of gross sales within 4 months
  • Built Loan Against Mutual Funds from zero — ₹100Cr disbursals in 7 months
  • Led ET Money Earn from 0 to ₹300Cr AUM
  • Built execution-only Mutual Fund Platform — projected ₹10Cr annual revenue
Ask my agent about the support agent eval pipeline at ET Money →

Projects

Krux.news

AI news product that turns fast-moving AI updates into short, swipeable stories.

  • Built the content pipeline and consumer web app.
  • Designed a multi-stage AI workflow from ingestion to publishing.
AI curationWeb researchPublishing
Krux.news product screenshot

MF Search

Semantic search engine across 16,197 mutual funds with intent-aware routing.

  • Routes queries into lookup, comparison, or filtered discovery.
  • Handles typos, partial names, categories, and investment styles.
  • Generates fund analysis across returns, risk, and expense ratios.
EmbeddingsIntent routingChromaDB

US Stocks Agent

Multi-agent stock analysis system with financial tools and voice input/output.

  • Improved tool-use accuracy from 58% to 90%.
  • Moved from one agent to specialist handoffs after evals.
  • Covers 20+ test cases across 5 query categories.
Agent handoffsVoice AIEvals

Feedback Intelligence

Cross-channel agent that turns public feedback into prioritized product signals.

  • Ingests Play Store, App Store, YouTube, and X feedback.
  • Classifies bugs, UX complaints, feature requests, and sentiment trends.
  • Separates UI chat state from persistent model session memory.
Agents SDKFunction toolsMemory

Also built

  • TalkToKrishnaRAG product grounded in 700 Bhagavad Gita verses with a 50-question eval suite. README ↗ / Know More ↗
  • Evaluation FrameworkLLM-as-judge and voice evaluation system for quality, safety, and ASR accuracy. README ↗ / Know More ↗
More on GitHub ↗

Education

MBA (Executive), SP Jain Global School of Management

2025–Present

BBA, Finance & International Business, Christ University

2017–2020

Summer School, Alternative Investments, London School of Economics

2018

If you have any more questions