Anil Prasad

28 years building production AI/ML at Fortune 100 scale. Founder @AmbhariiLabs. Head of Engineering & Product @DukeEnergy. I write about what actually breaks in regulated AI deployments.

Closter, NJ · New York Metro Joined on Mar 25, 2026 https://ambharii.com

Education

Stanford University · BITS Pilani

Pronouns

He/Him

Work

Head of Engineering & Product · Duke Energy

Anil Prasad

Sep 19

An agent harness where stopping is scored and the transcript is not the agent's to write

#ai #python #opensource #security

8 min read

Anil Prasad

Sep 12

Agent Handoff Contracts: A CI Gate for Multi-Stage Pipelines

#ai #python #tutorial #devops

7 min read

Anil Prasad

Aug 23

Instrument the bill, not just the model

#discuss #ai #mlops #architecture

4 min read

Anil Prasad

Aug 14

Put a scoring gate in front of your LLM call, not a human

#ai #python #architecture #mlops

6 min read

Anil Prasad

Aug 7

Measure pass-all-k, not accuracy: a reliability harness in 60 lines

#ai #python #mlops #testing

5 min read

Anil Prasad

Aug 1

I Have 25+ Years of Production Engineering Experience. Here Is What AI Coding Tools Actually Did to My Workflow.

#ai #productivity #programming #career

7 min read

Anil Prasad

Jul 26

Your model didn't get smarter. It learned to cheat the test.

#machinelearning #llm #python #opensource

3 min read

Anil Prasad

Jul 21

7 New Governance Features Just Shipped in AgentMesh. All Open Source, No Waitlist. published: false

#ai #opensource #security #eu

6 min read

Anil Prasad

Jul 10

Building Production-Safe AI Agents in 2026: A Governance-First Architecture

#ai #python #tutorial #opensource

5 min read

Anil Prasad

Jul 1

Title: I Built an AI Governance Proxy in 72 Hours. Here Is Exactly How.

#ai #opensource #python #security

6 min read

Anil Prasad

Jun 26

I built a zero-dependency PII scanner for AI prompts in 270 lines of Python

#ai #python #security #opensource

5 min read

Anil Prasad

Jun 14

I put one proxy in front of every AI tool my team uses 85% cache hits, 75% lower cost

#ai #machinelearning #python #opensource

4 min read

Anil Prasad

Jun 12

How I took a production RAG pipeline from 61% to 97% accuracy (6 stages, full code)

#ai #machinelearning #python #anilprasad

5 min read

Anil Prasad

Jun 1

How We Cut AI Infrastructure Costs by 94% Without Sacrificing Quality (And How You Can Too)

#ai #machinelearning #llm #aiops

13 min read

Anil Prasad

May 19

Building Production-Ready Open Source AI Infrastructure: A Technical Guide

#opensource #datascience #ai #machinelearning

7 min read

Anil Prasad

May 14

How We Cut LLM API Costs by 94%: A 3-Layer Caching Strategy

#ai #performance #productivity #tutorial

5 min read

Anil Prasad

May 12

Building Production RAG: From 52% to 89% Accuracy with a 6-Stage Pipeline

#ai #python #machinelearning #tutorial

5 min read

Anil Prasad

May 8

The web is now weaponized against your AI agents

#productionai #llmops #agentsecurity #aigovernance

5 min read

Anil Prasad

May 6

Claude Code Has 6 Ways to Authenticate. I Built a Cross-Platform Installer Because of It

#devops #ai #opensource #claude

4 min read

Anil Prasad

May 2

The week the agent capability inflection arrived. And what to do about the 86% that still fail.

#ai #agents #machinelearning #aiops

6 min read

Anil Prasad

Apr 24

The week AI capability outpaced readiness. Again. Here is what it means in production.

#ai #argus #genomixiq #ariarcm

5 min read

Anil Prasad

Apr 18

I Built the First Agentic AI Platform for Clinical Genomics. Here Is the Full Architecture

#agentaichallenge #genomics #healthtech #ai

7 min read

Anil Prasad

Apr 10

I Built 11 Autonomous Agents for Healthcare Revenue Cycle. Here Is the Full Architecture.

#ai #python #machinelearning #healthcare

9 min read

Anil Prasad

Apr 8

Building Multi-Agent Systems That Don't Collapse in Production

#ai #agentaichallenge #machinelearning #powerplatform

5 min read

Anil Prasad

Apr 7

How I Built a Self-Correcting Multi-Agent System for Healthcare - and Why Standard ML Metrics Failed Me

#ai #python #machinelearning #healthcare

9 min read

Anil Prasad

Apr 4

78% of PyTorch Models Never Reach Production. I Built the Fix.

#pytorch #mlops #opensource #python

6 min read

Anil Prasad

Apr 2

I open sourced a production MLOps pipeline. Here is what it took to get it to PyPI and Hugging Face in one day.

#mlops #python #opensource #machinelearning

4 min read

Anil Prasad

Mar 28

argus-llm is now on PyPI — production LLM observability in one pip install

#llm #ai #machinelearning #aiobservability

1 min read

Anil Prasad

Mar 25

OpenTelemetry Traces Your LLM. It Does Not Fix It.

#llm #observability #opensource #ai

3 min read

DEV Community

Anil Prasad

Badges

Writing Debut

Skills/Languages

Currently learning

Currently hacking on

Available for

An agent harness where stopping is scored and the transcript is not the agent's to write

Agent Handoff Contracts: A CI Gate for Multi-Stage Pipelines

Instrument the bill, not just the model

Put a scoring gate in front of your LLM call, not a human

Measure pass-all-k, not accuracy: a reliability harness in 60 lines

I Have 25+ Years of Production Engineering Experience. Here Is What AI Coding Tools Actually Did to My Workflow.

Your model didn't get smarter. It learned to cheat the test.

7 New Governance Features Just Shipped in AgentMesh. All Open Source, No Waitlist. published: false

Building Production-Safe AI Agents in 2026: A Governance-First Architecture

Title: I Built an AI Governance Proxy in 72 Hours. Here Is Exactly How.

I built a zero-dependency PII scanner for AI prompts in 270 lines of Python

I put one proxy in front of every AI tool my team uses 85% cache hits, 75% lower cost

How I took a production RAG pipeline from 61% to 97% accuracy (6 stages, full code)

How We Cut AI Infrastructure Costs by 94% Without Sacrificing Quality (And How You Can Too)

Building Production-Ready Open Source AI Infrastructure: A Technical Guide

How We Cut LLM API Costs by 94%: A 3-Layer Caching Strategy

Building Production RAG: From 52% to 89% Accuracy with a 6-Stage Pipeline

The web is now weaponized against your AI agents

Claude Code Has 6 Ways to Authenticate. I Built a Cross-Platform Installer Because of It

The week the agent capability inflection arrived. And what to do about the 86% that still fail.

The week AI capability outpaced readiness. Again. Here is what it means in production.

I Built the First Agentic AI Platform for Clinical Genomics. Here Is the Full Architecture

I Built 11 Autonomous Agents for Healthcare Revenue Cycle. Here Is the Full Architecture.

Building Multi-Agent Systems That Don't Collapse in Production

How I Built a Self-Correcting Multi-Agent System for Healthcare - and Why Standard ML Metrics Failed Me

78% of PyTorch Models Never Reach Production. I Built the Fix.

I open sourced a production MLOps pipeline. Here is what it took to get it to PyPI and Hugging Face in one day.

argus-llm is now on PyPI — production LLM observability in one pip install

OpenTelemetry Traces Your LLM. It Does Not Fix It.