$ itssujeeth — Principal Applied AI Architect
I am an enterprise AI architect and hands-on technology leader with 24+ years across software engineering, enterprise architecture, cloud platforms, and applied AI. I help Fortune-class organizations turn LLMs, RAG, agentic workflows, and AI platforms into production-ready systems that integrate with real enterprise data, identity, security, and operating models. My lens is practical: separate durable engineering patterns from AI noise, then turn them into architectures, specs, prototypes, and adoption roadmaps teams can actually use.
Launched 3 flagship production RAG applications within 18 months by translating enterprise GenAI blueprints into funded, production-scale AI products.
▰ Picks
Independent, unsponsored. Conviction rated 1–5.
The fastest path from intent to shipped code I've found. What makes it different from every other AI coding tool is agentic mode with real persistent context — it finishes tasks, not just suggests lines. If you're building in this ecosystem and aren't using it daily, you're leaving hours on the table. Non-negotiable.
npm install -g @anthropic-ai/claude-code The cleanest way to route across Claude, GPT-4o, Gemini, and local models behind a single OpenAI-compatible endpoint. Saves you from hardcoding provider SDKs everywhere and makes switching models a config change, not a refactor. The proxy mode is the real value — run it once, point all your agents at localhost.
pip install litellm Solves the stale-docs problem that plagues AI coding assistants. Pulls live, version-accurate library documentation directly into Claude's context so you stop getting hallucinated API signatures. If your team ships against fast-moving SDKs — Astro, LangChain, MCP itself — this is a quiet force multiplier.
npx -y @upstash/context7-mcp The friendliest way to run local models — one command, no Python environment drama. Indispensable for offline experiments, cost-free eval loops, and keeping sensitive data off third-party APIs. Honest caveat: local model quality still trails frontier models on complex reasoning, so treat it as a complement, not a replacement.
brew install ollama The most useful MCP server directory right now — good breadth, install snippets, and a growing catalog. Valuable as a starting point for discovery; do your own due diligence before wiring any server to a production agent. Quality and maintenance levels vary significantly across listings.
High install count, heavy abstraction layer, and a habit of hiding the actual LLM call from you when you most need to see it. Good for getting a RAG prototype running in an afternoon; not where I'd want to be when debugging a production agent at 2am. Worth understanding the patterns it encodes — not the dependency itself.
▰ Selected work
All projects →An LLM-driven pipeline that automatically remediates known security vulnerabilities across an enterprise codebase, replacing slow manual patching.
Three production retrieval-augmented-generation applications giving non-technical staff natural-language access to enterprise data and decision support.
A five-zone enterprise agentic-AI platform that lets non-technical business teams self-serve AI-generated insights from governed enterprise data across six business domains.
A single command that points Claude Code at any inference backend — local Ollama, Anthropic, OpenAI, or Gemini — with a clean environment and no manual env-var juggling.
A 24-week, self-paced interactive curriculum — shipped as a React app — that teaches IT support managers to use Claude across their actual daily work.
▰ Career
Lead enterprise AI, cloud, data, and agentic platform architecture for Fortune-class clients, translating emerging AI platform capabilities into practical adoption roadmaps and production patterns.
Drove enterprise GenAI adoption, AI reference architectures, automation strategy, and production AI delivery within a large logistics enterprise.
Led AI/ML product architecture, portfolio rationalization, secure platform design, observability standards, and innovation initiatives.
Designed and delivered enterprise-scale software, cloud, workforce, logistics, and retail technology solutions for Walmart and other large clients.
Led architecture, design, coding, testing, integration, and regulatory delivery across pharmacy, tax, medical telemetry, and enterprise systems.
Built and maintained commercial software products across tax processing, backup systems, EMR, and business applications.
Started career in software development, training, lab operations, and internal automation systems.
Delivered early-career business automation software for small businesses and local organizations.
▰ Writing