Sandhu Santhakumar

Sandhu is a Technical Architect specializing in scalable cloud systems, distributed architectures, and microservices-based systems. With deep hands-on experience in Java, Python, AWS, and modern DevOps practices, he designs and delivers high-performance systems that balance scalability, reliability, and real business impact. His work spans distributed systems design, integrating AI/ML capabilities into production systems, and architecting platforms built for long-term growth—not just short-term delivery.

Prompts Are Code Now: Versioning, Testing, and Shipping

The Prompt That Quietly Broke Production A one-line prompt change shipped on Friday. By Monday, your support bot is confidently telling customers that refunds take 3 days, even though the policy changed to 7. No alerts fired, no errors logged, and no crashes. Just silent, scalable misinformation. This is the new class of production bug: […]

Read more →
Slashing LLM Costs: Why We’re Moving to TOON for Heavy Payloads

Slashing LLM Costs: Why We’re Moving to TOON for Heavy Payloads

TL;DR We’re spending too much on JSON overhead in LLM calls. By switching to TOON (Token-Oriented Object Notation) at the LLM boundary only, we can cut token usage by 40–60% for structured data without sacrificing accuracy, compatibility, or developer sanity. Last month, while reviewing our AWS bills for a marketing automation project, I noticed a […]

Read more →