
Preetam Ramdhave
I deploy AI where it has to work.
Architecting resilient, agentic AI systems and cloud-native platforms for Fortune 500s and high-growth scaleups. 18 years of enterprise experience. Based in Seattle.
Selected Outcomes — Real Customers, Real Numbers
Shipping production systems end-to-end — UK enterprise clients, Fortune 500, solo founder products.
Manual review effort eliminated via the org's first production agentic AI workflow.
Handwritten prescriptions processed in 48 hours at a charitable healthcare event.
"Delighted Customer" awards (2012–2014) voted by UK enterprise clients for direct impact.
Featured Deployments
Full-Stack · Agentic AI · Healthcare
Open Source & Free Apps
Step 06 · Generalize — turning one customer's win into a reusable pattern.

NervaPack
91.2% token reduction · 96% recallPrivacy-first, 100% offline knowledge graph & bi-temporal memory engine for AI coding agents. Uses Tree-sitter AST parsing, K-hop BFS graph traversal, and SQLite FTS5 bi-temporal facts storage to cut token consumption by 91.2% while achieving 96% recall on SWE-bench Lite. Features a dual MCP server suite (20 tools across code graph and memory) native for Claude Code, Cursor, and Windsurf.
OmmSai — Healthcare AI Pipeline
15K PDFs · 20× fasterOpen-source Python pipeline I donated to a charitable healthcare event that needed to digitize 15,000+ handwritten prescription PDFs in 48 hours. Built with Claude Sonnet (Anthropic API), Google Drive API, ThreadPoolExecutor concurrency, and a Tkinter operator GUI so non-engineer volunteers could run it on any Windows laptop. Released publicly so other charities and clinics can reuse the pipeline.
Handwriting JSON
Published on PyPIOpen-source Python package and CLI for automating handwritten document workflows. Converts forms, notes, and scanned paperwork into structured JSON using vision LLMs and optional schema guidance.
timesfm-mcp
Published on PyPI · v0.1.6An MCP server that gives any AI agent zero-config time-series forecasting. Plugs Google's TimesFM 2.5 foundation model — plus a pure-NumPy statistical baseline — into Claude, Cursor, or any MCP client via one install line. The agent calls a forecast tool, gets point predictions, confidence bands, and a trend/seasonality summary, then writes the recommendation itself. Includes a backtest tool reporting MAE/sMAPE on held-out data so forecasts are validated before they're trusted.

JapaApp
Free · LiveSpiritual mantra-tracking PWA built for practitioners of Vedic disciplines. Originally architected on AWS (Lambda, RDS Proxy, Cognito JWT, SAM) before owning the platform-migration decision to Firebase. Implemented global admin via Firebase Custom Claims and a free-to-use experience with an optional Razorpay donation flow.
How I Work
The Forward Deployed Engineer motion — same pattern across UK enterprise clients, Fortune 500, and solo founder products.
Embed
Sit with the customer — external client, internal department, or end user. Watch how they actually work.
Discover
Find the real problem. It is almost never the stated problem.
Design
Architect the solution that fits the customer's reality — their data, their systems, their team, their security posture.
Ship
Build it end-to-end. Backend, frontend, infra, security, observability. No handoffs.
Operate & Iterate
Stay with it after launch. Watch the customer use it. Iterate on what the field teaches.
Generalize
Turn one customer's win into a reusable pattern the rest of the org can leverage.
Recent Writing
Read all articlesField Notes
Field notes from production AI deployments. Monthly. Short.
No spam. Just code, AI, and entrepreneurship. Unsubscribe anytime.
