building with AI

Deepak Kaushish Sharma

I build AI products end-to-end from psychology-first principles. Four shipped products, all live in production. 15 years across product, engineering, online store and growth at OYO, Foodpanda, Delivery Hero, Zepto, Dell, and Citi. I don't just use AI, I build behavioral intelligence systems around it.

Python React Next.js Node.js AWS MongoDB LLMs Behavioral AI
Kritheon

Behavioral intelligence platform that replaces traditional market research with prediction.

Describe your product, campaign, or strategy and get a complete prediction report: how your target audience will respond, what drives their decisions, and where they'll push back.

Our behavioral intelligence profiles evaluate your concept independently, reporting back with purchase intent, emotional response, and detailed reasoning. Every prediction comes with full statistical validation, confidence intervals, and segment-level breakdowns. Results delivered in minutes, not the 6-8 weeks a research agency would take.

Tested and validated across India, US, Canada, UAE, and UK with multi-vertical coverage.

Live Python LLMs MongoDB Vercel
CinePredict

A virtual test screening for your script.

Upload your screenplay, get a complete greenlight package: audience intelligence, box office modeling, marketing direction, and talent impact analysis in one report.

Our intelligence profiles experience your script and report back individually with viewing interest, reasoning, and recommendations. The financial engine runs 10,000 Monte Carlo simulations with break-even and ROI at multiple budget levels. You also get specific trailer hooks, taglines, viral moments, and risk flags your marketing team can act on immediately.

Works natively for both Hollywood and Indian cinema with regional breakdowns.

Live React Python LLMs Three.js Clerk Auth
InstaRecs

AI recommendation engine that makes everyday decisions effortless.

Tell it what you're deciding and get a personalized recommendation backed by reasoning. Food, gifts, fashion, entertainment, and travel, all tailored to your preferences, context, and constraints.

The engine learns your taste profile across categories and sharpens its suggestions over time. Not generic "top 10" lists, but specific picks with explanations for why they fit you. Handles nuance like budget, dietary restrictions, occasion, and mood.

Expanding into investment guidance, book recommendations, and user-created decision modules.

Live Python LLMs Flask MongoDB AWS
Kidsightful

AI-powered child development platform that gives parents real clarity about their kids.

Complete a structured, age-specific assessment and get a personalized developmental profile: cognitive patterns, emotional tendencies, social dynamics, and learning style mapped across key growth dimensions.

The intelligence engine analyzes responses against developmental benchmarks for ages 2-18, generating actionable guidance tailored to your child's unique profile. Not generic parenting advice, but specific insights a child psychologist would take multiple sessions to surface.

Covers the full developmental spectrum from toddler milestones to adolescent behavioral patterns.

Live Next.js LLMs Clerk Auth Stripe
Stack
Full-stack. Python backends, React/Next.js frontends, cloud-native on AWS.
AI
LLMs, behavioral modeling, predictive systems. Not wrappers, real AI architecture.
Shipping
All four products are live, in production, serving real users.
Solo
Designed, built, and deployed every product end-to-end. No team, no outsourcing.
VP Technology
Leading Global Financial Institution
Technology strategy and business analysis at scale
Head of Catalog
Zepto
High-intensity sprint fixing catalog systems at India's fastest-growing quick commerce company
Director, Content Ops
OYO
Scaled content operations at India's largest hospitality company
Global Head of Product
Delivery Hero / Foodpanda
Owned product globally for the logistics and online store platform
15+ years in Product & Engineering
Across consumer tech, fintech, and AI