Raashi Shah
Technical Product Manager, AI Engineer
Designing and developing apps and AI agents end-to-end
Admissions Evaluation Agent
Academic institutions process thousands of applications every year
I standardised this workflow with enterprise-grade agents made with Google's ADK
- Created tools to work with bad data and rank it consistently
- Works with past data for insights
Used telemetry to measure and bring cost down to 15 cents per student
Expression - Animation Tool
Auto-colouring frames hand drawn by an animator, in a style called frame-by-frame animation
This problem remains unsolved worldwide
Frames are coloured one by one, for 1 min, 25 fps shots that's 1,500 frames
My goal is for artists to retain full control while giving them better tools
Side quest: first ever to parse line art in PNG files 1:1, as the artist intended, it works with any software the artist uses
Offline Expo Navigation
Web app for an architecture exhibition, Design POV
Works offline regardless of footfall
Made 99% of this in 12 hours
Pluto
2021-2024
Collaborated cross-functionally with artists and transformed a creative studio into a product and tech-led team
Launched 3 digital asset products:
Magic Batch - MVP
First ever project in the ecosystem to introduce cross-platform payments, driving 37% sales growth YoY
Increased retention by 82% through A/B-tested incentives
500+ users generated 5k+ digital assets within 10 days, including some made with image-gen models
Kotak Securities
2021
Short stint, founding hire on Kotak Neo's product team - consumer trading app
Kawa Space
Geospatial data analysis using machine learning in agriculture, rainfall, and population density
Created a chatbot using Dialogflow for users to query data in natural language just by asking, this was pre-GPT-3
Aula Education
Led customer success and engagement analytics for the team's largest portfolio, expanding from 1 to 3 universities across the UK and the US
Boosted retention from 9.2% to 32% by designing a data-driven analytics toolkit that combined qualitative insights with quantitative usage data
Doubled engineering delivery speed by analysing user feedback and partnering with VP of Product to implement agile workflows