Back to Products
Siramai
View DemoArchived
AI/ML
Agentic e-commerce OS for 4+ enterprise brands—2K+ daily queries, 30% latency drop, $1.0M raised.
4+Customers
2K+Daily Queries
30%Latency ↓
Problem
E-commerce brands lose conversions due to poor keyword-only search that can't personalize at scale.
Solution
No-code agentic framework with multimodal vector search and adaptive LLM reranking.
Impact
4+ enterprise customers, 2K+ daily queries, 30% latency reduction, 18% relevance improvement. $1.0M raised.
Why I Built This
E-commerce brands were losing customers to poor search. Built this to make AI search accessible without ML expertise.
Architecture
Architecture
Next.js
Vector Search
Gemini
PostgreSQL
Redis
Kubernetes
AWS EKS
Technical Highlights
1
Designed no-code agent framework + adaptive ranking: vector search + LLM reranking → 30% latency reduction at 2K+ QPS
2
Deployed on AWS EKS with Redis caching + auto-scaling—99.9% uptime across 4+ enterprise customers
3
Optimized multimodal search pipeline (text + image embeddings) for e-commerce catalog indexing, enabling real-time personalization
Technologies & Tags
ML/Agents
Vector Search
Infra