2023 - Ongoing
Lead Engineer
CookingAI
An AI-driven recipe intelligence platform that optimizes personalization, nutrition insights, and supplier integration.
Problem
CookingAI needed to scale from a curated beta to a global marketplace while maintaining culinary quality.
The existing system struggled with personalization, nutrition logic, and real-time supplier updates.
Approach
Designed a modular AI pipeline with deterministic scoring for nutrition and preferences.
Introduced a feature store to unify user taste signals across web and mobile touchpoints.
Architecture
Hybrid retrieval and ranking services using vector search for recipe similarity.
Event-driven ingestion for suppliers with validation gates and rollback safety.
Impact
Improved conversion from recipe view to cart by 21%.
Enabled localized cooking experiences across 12 regions with shared infrastructure.
Stack
Next.js, Node.js, Python, OpenAI, Redis, PostgreSQL, AWS Lambda, Terraform.
Learnings
Operational reliability requires both AI guardrails and strong data contracts.
Small latency wins compound into a drastically better user experience.
Impact
- ●Reduced recipe recommendation latency by 43% through caching and ranking optimization.
- ●Launched multilingual recipe assistant with 98% uptime in the first 6 months.
- ●Built supplier ingestion pipeline processing 120k ingredient updates per day.