Rithik Pyne
Backend engineering, Python, and AI systems. Notes on building reliable software with language models.
2026 2 posts
- Mar Building Agentic AI Systems Beyond Simple Chatbots A practical engineering view of agents as systems that plan, use tools, manage state, and recover from failure.
- Jan Observability for AI Systems: Beyond 200 OK How backend and AI engineers monitor prompts, retrieval, tool calls, latency, cost, safety, and output quality in real production systems.
2025 3 posts
- Mar Evaluating LLM Applications: Why Unit Tests Are Not Enough How engineers test probabilistic outputs, retrieval quality, regressions, safety, latency, and cost in real LLM systems.
- Feb RAG in Production: Retrieval, Chunking, Embeddings, and Evaluation How real RAG systems handle ingestion, chunking, retrieval quality, context construction, evaluation, latency, and stale data.
- Jan API Rate Limits, Cost, and Latency in LLM-Powered Systems How backend and AI engineers handle rate limits, token usage, latency, retries, caching, fallbacks, and cost control when building with external model APIs.
2024 5 posts
- Nov What AI Engineering Looks Like in Practice Turning models, prompts, tools, and data into reliable production systems.
- Sep Backend Engineering for High-Volume Data: Queues, Batching, and Backpressure How backend systems stay reliable when requests, events, and data pipelines move faster than the application can process them.
- Jul Building Reliable Python APIs: Timeouts, Retries, Idempotency, and Observability How production Python services handle unreliable networks, duplicate requests, slow dependencies, and failures you cannot avoid.
- May Designing Scalable System Architectures for Real Workloads A practical guide to designing backend systems around workload patterns, bottlenecks, data flow, and failure modes.
- Mar Python Concurrency in Practice: When to Use Threads, Processes, or Asyncio A practical guide to choosing between multithreading, multiprocessing, and async code in real backend systems.