Ayaan Khan

1. Designed a 16-service polyglot backend (Python/FastAPI, Go/gRPC) on AWS EKS supporting multi-framework AI agent execution — processed 341 agent runs and 12,455 billing events within 24 hours of launch.

2. Engineered a 3-tier checkpoint system (Redis → PostgreSQL → S3) achieving <5ms write latency; added circuit breaking, DB-fallback on gRPC outage, and async LZ4-compressed S3 offloading.

3. Built a self-healing RLHF pipeline with MAST failure classification, ELO-ranked multi-LLM fix generation, and automated GitHub PR creation — enabling AI-driven recovery without human interventio

Building a platform to master Linux CLI and Bash using real-world, interactive terminal challenges.

Architecting a RAG system to query CRM, Excel, and Salesforce data using Pinecone and Knowledge Graphs Neo4j. • Designed modular AI pipelines for scalability, integrating vector search and Graph RAG.