Randy Tatham

Hello, I'm

Randy Tatham


Technical Executor and Leader with over 15 years of experience building and optimizing teams on government programs through development of data-focused scalable processes and technical solutions. Successfully executed programs in excess of $100M+ as an individual contributor, technical lead, and program manager.

/ Career

A brief overview of my career

April 2026 - Present

Director, Foundations Geo Expansion & AI Transformation

Microsoft

  • Managing an 11-person engineering organization driving strategic planning and cross-functional coordination.
  • Led OKR creation and quarterly planning cycles.
  • Designed a Large Language Model–based automated judging system reducing manual review time by 60% while improving scoring accuracy.
Engineering Leadership LLM / AI OKR Planning IL6 Cleared
Jan 2026 - April 2026

Director of Core Platform & Services

SynthBee

  • Managed an 11-person engineering organization driving strategic planning and cross-functional coordination.
  • Led OKR creation and quarterly planning cycles.
  • Designed an LLM-based automated judging system reducing manual review time by 60% while improving scoring accuracy.
Engineering Leadership LLM / AI OKR Planning Strategic Planning Python
Feb 2022 - Jan 2026

Power Platform Principal Engineering Manager

Microsoft

  • Managed 24 employees delivering Power Automate, Power Apps, and Dataverse into air-gapped government clouds.
  • Achieved 54% improved service parity across public and air-gapped clouds.
  • Drove a 70% improvement in software delivery timelines through Runbook development.
Power Automate Power Apps Dataverse Azure Air-gapped Clouds Engineering Management
Mar 2006 - Feb 2022

Mission Control Systems CI/CD Manager & Technical Lead

Lockheed Martin

  • Drove ADP digital transformation across a $1.2B line of business.
  • Technical Author on 6 proposals resulting in $58M in acquisition.
  • Achieved 75% defect capture rate improvement.
  • Architected interoperability solutions for heterogeneous UAV and GCS systems across multiple classified programs.
  • …and approximately 847 other things that would require security clearance to explain. 🔒
Java Python C# UCI CI/CD Mission Systems System Engineering Program Management

/ Agentic Transformation

Leading the shift from tools to autonomous AI teammates

Engineering organizations that embrace agentic AI don't just move faster — they think differently. I architect and lead teams that integrate AI agents as first-class contributors, transforming how software gets designed, built, reviewed, and shipped.

Agents as Team Members

AI agents — Copilot, Squad, Claude Code — are not autocomplete. They are autonomous contributors with defined roles, memory, and accountability. I design organizations where humans and agents collaborate with clear handoffs.

Agentic Workflows

From issue triage to PR review to deployment, I build pipelines where agents handle routine work autonomously — freeing engineers to focus on architecture, product decisions, and creative problem-solving.

Measurable Impact

Agentic transformation is not abstract. I track cycle time reduction, defect escape rates, and engineering velocity — delivering 60–75% improvements in key delivery metrics across programs.

Human-AI Upskilling

The hardest part of agentic transformation is cultural. I coach engineering teams to think in agents, prompts, and workflows — making AI augmentation a sustainable competitive advantage, not a one-time experiment.

Tools & Platforms

GitHub Copilot Copilot CLI Claude Code Squad GitHub Copilot for Teams MCP Cursor Aider Copilot Studio Azure AI

/ Featured Projects

A selection of professional work, published apps, and experimental prototypes.

🔒 Classified
AI System

LLM-Based Automated Judging System

Designed a Large Language Model–based automated judging system enabling consistent, scalable evaluation of submissions. Reduced manual review time by 60% while improving scoring accuracy across multiple domains.

Python LLM AI Automation
🔒 Classified
Defense System

Cache: Service Bus Consumer

Multi-threaded, multi-application-service-bus consumer written in Java that stores compressed information enabling post-mission debrief and analysis of distributed UCI systems.

Java Multi-threading Service Bus UCI
🔒 Classified
Research

Interactive Learning Mission Planner

Applied interactive learning to achieve 80% improvement in automated aircraft route creation. Architected a composable system through Service Oriented Architecture (SOA) design patterns with automated test event metrics.

Python Machine Learning SOA Pandas Matplotlib

/ Tech Stack

Breakdown of my technical skills, categorized by domain

Programming Languages

Python
C#
JavaScript

Microsoft Power Platform

Power Apps
Power Automate
Copilot Studio
Dataverse

Cloud & Azure

Azure
Azure Data Explorer
Cosmos DB
Azure SQL
Service Fabric

DevOps & Tools

JIRA
Confluence
Docker
Git
Azure DevOps
GitHub Actions

Methodologies

Agile / Scrum
Scaled Agile (SAFe)
Kanban
OKR Planning

AI Agents & Agentic Tools

GitHub Copilot
Copilot CLI
Claude Code
Squad (AI Teams)
GitHub Copilot for Teams
Cursor
Aider

AI Engineering

Prompt Engineering
LLM Integration
MCP (Model Context Protocol)
RAG Pipelines
AI Workflow Automation

/ Contact Me

Feel free to reach out. I'm always open to discussing new projects and opportunities.