Ajay Varma Tanniru

About

Backend engineer who built an e-commerce order processing service in Python/Flask with PostgreSQL, focusing on correctness and reliability (idempotency, Redis locks, async payment processing with circuit breakers). Also integrated an ML recommendation system as a separate FastAPI inference service with caching and async embedding updates, reporting ~25% CTR lift, and has experience with multi-tenant isolation using PostgreSQL row-level security.

Experience

Education

Key Strengths

• Designed ACID-safe order processing and payment tracking with PostgreSQL
• Prevented duplicate orders/double charging using idempotency keys, Redis locks, and DB unique constraints
• Improved reliability of payment integrations via Celery async processing, retries with exponential backoff, and circuit breaker
• SQLAlchemy/PostgreSQL performance tuning (indexes, composite indexes, partitioning, avoiding N+1 with selectinload/joinedload, raw SQL for heavy aggregations)
• Production AI/ML integration: model training (TensorFlow/PyTorch) and serving via FastAPI/Flask with Docker/Kubernetes
• Delivered measurable ML impact: ~25% click-through rate increase from e-commerce recommendations with caching and async updates
• Implemented multi-tenant data isolation using tenant_id scoping and PostgreSQL row-level security
• Scaled high-throughput background processing by batching tasks, caching, autoscaling workers, and adding retries/rate limiting