Skip to content
View abh2050's full-sized avatar

Block or report abh2050

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
abh2050/README.md

Abhishek Shah, PMP

AI Engineer & Technical Program Manager Enterprise AI · Multi-Agent Systems · AEC Industry

📍 Portland, OR  |  SSOE Group  |  MS AI/ML, University of Michigan

LinkedIn · Email · Medium · Podcast


About

I build production multi-agent AI systems for enterprise environments where compliance, governance, and reliability are non-negotiable. Currently governing the Abacus AI enterprise platform at a top-ten AEC firm, managing a Microsoft Copilot rollout across hundreds of engineers, and running a vendor oversight framework for external RAG and multi-agent workstreams.

Before this role: Started as a chemical engineer, then pivoted to AI while at Intel. At BMW of North America, shipped the AI Champions program across manufacturing plants. I think in systems, operate at the intersection of technical depth and organizational change, and write about it on Medium.


What I Build

01 · Production Agent Systems Orchestrated sub-agent architectures with state machines, memory persistence, and harness-level reliability engineering. LangGraph, Abacus AI, Anthropic SDK.

02 · Enterprise RAG Pipelines Page-indexed Azure RAG for large document corpora, multi-source evidence reconciliation, and hybrid retrieval across heterogeneous data sources.

03 · AI Platform Governance AI App Store frameworks, vendor workstream management, enterprise LLM rollouts. Security-first: Microsoft Defender, SentinelOne, compliance-grade logging.

04 · AI Program Management Technical roadmaps, stakeholder alignment, and vendor evaluation at the enterprise scale. PMP-certified. Built for organizations where AI is infrastructure, not experiment.


Featured Projects

Pathfinder — Multi-Agent Sales Intelligence System ACTIVE Production multi-agent architecture with specialized sub-agents: market discovery, account intelligence, competitive intelligence, pre-RFP enablement, and portfolio learning. Evaluating Anthropic Managed Agents API vs Agent SDK for enterprise compliance. Built for AEC sales cycles.

Abacus AI Enterprise Platform — SSOE Group ENTERPRISE Governing the full Abacus AI deployment across a 2,000+ person AEC firm. AI App Store governance framework, vendor workstream oversight, multi-agent and RAG system delivery, and Microsoft Copilot rollout program management.

Azure Page-Indexed RAG Architecture RESEARCH Designed for large-scale enterprise document collections: page-level indexing, hybrid retrieval, multi-source reasoning with parallel orchestration and evidence reconciliation across star-schema heterogeneous sources.

BMW AI Champions Program SHIPPED Designed and oversaw the global AI Champions program at BMW of North America, collaborating with manufacturing plants in Germany and China. Drove cross-functional AI adoption in a highly regulated automotive production environment.


Tech Stack

Domain Tools
AI Platforms Abacus AI, Anthropic SDK, OpenAI, AWS Bedrock, DataRobot
Agent & RAG LangChain, LangGraph, Weaviate, MCP, DSPy
Cloud & Infra Azure, AWS, Docker, Kubernetes, GCP
Data & ML PySpark, XGBoost, TensorFlow, Airflow, PostgreSQL
Languages Python, SQL, Bash
Security Microsoft Defender for Endpoint, SentinelOne, XDR

Technical Competency

Chemical engineer turned AI architect. I bring the rare combination of technical depth, enterprise program management, and end-to-end delivery at scale.

Connect on LinkedIn

Pinned Loading

  1. financial_doc_analyser financial_doc_analyser Public

    This tool automates the retrieval and processing of SEC filings (e.g., 10-K, 10-Q, 8-K) using a company's ticker symbol. It intelligently converts PDFs and HTML documents into a structured vector d…

    Python 1 1

  2. financial_doc_analyser_local_ollama_Deepseek financial_doc_analyser_local_ollama_Deepseek Public

    A Streamlit-based application designed to download, process, and analyze SEC filings (e.g., 10-K, 10-Q, 8-K). This tool leverages AI-powered embeddings, vector databases, and advanced document proc…

    Python 2

  3. hybrid_rag_with_bm25_reranker hybrid_rag_with_bm25_reranker Public

    This project implements an Advanced Hybrid Retrieval-Augmented Generation (RAG) System using Streamlit. It allows users to upload documents (TXT, PDF, DOCX), processes the content into manageable c…

    Python

  4. langgraph_multi_agent_ai_travel_agent langgraph_multi_agent_ai_travel_agent Public

    A state-of-the-art multi-agent travel planning system built with LangGraph, Google Gemini Flash-2.0, and DuckDuckGo Search. The system offers three different planning approaches, from traditional s…

    Python 5 4

  5. pump_predictive_maintenance_using_Markov_decision_gemini pump_predictive_maintenance_using_Markov_decision_gemini Public

    This app does pump predictive maintenance using Markov Decision process and Gemini API

    Jupyter Notebook

  6. customerservice_swarm_agent customerservice_swarm_agent Public

    This project implements a multi-agent system (Agent Swarm) that processes user requests through specialized agents working together. The system includes a Router Agent, Knowledge Agent with RAG cap…

    Python 4