Ratish Jain
AI Engineer
Over the past few years, I’ve worked across API design, distributed systems, DevOps, and AI workflows, helping take products from idea to production. I enjoy solving engineering problems end-to-end—whether that involves designing system architecture, building reliable backend services, automating infrastructure, or integrating AI where it creates real value. Interests: Backend Engineering, Platform Engineering, Cloud Infrastructure, Distributed Systems, and Applied AI.
{ Work Experience }
{ Certifications }
{ Recent writing }
I write about cloud architecture, AI engineering, and the occasional deep dive when something breaks in an interesting way.
Agentic Systems Are Just Patterns
Single Agent Patterns Tool Use Pattern An LLM by itself can only produce tokens. It can't query a database, hit an API, run code, or check today's weather. The tool use pattern closes that gap by givi
Aug 22, 2026 · 4 min read
Deploying a CrewAI Multi-Agent Crew to Amazon Bedrock AgentCore
Full source for this project: https://github.com/ratishjain12/crewai-bedrock-agentcore Amazon Bedrock AgentCore is AWS's answer to a problem every team building agents eventually hits: your agent wo
Aug 16, 2026 · 7 min read
LLM Observability: A Technical Deep Dive
How do you debug an LLM application when the answer is wrong, latency suddenly spikes, or token costs explode? Traditional application monitoring isn't enough. LLM applications require a new layer of
Jul 19, 2026 · 5 min read
{ Checkout my latest work }
I've worked on a variety of projects, from simple websites to complex web applications. Here are a few of my favorites.
GPT2Bench: Inference Engine from scratch and Benchmark
A CPU inference engine for GPT-2 (124M) implemented entirely from scratch in C, no PyTorch, no llama.cpp, no framework of any kind. Every component, tensor operations, the byte-level BPE tokenizer, multi-head attention, and the full transformer forward pass, was hand-written and verified byte-identical against Hugging Face's reference implementation. Ships with a live FastAPI + React dashboard for benchmarking two hand-built performance optimizations (a KV cache and multi-threaded matrix multiplication) in real time.
