Michael
Legemah
SaaS · AI Systems · Full-Stack
I build full-stack software solutions and production-grade AI systems. From RAG pipelines and LLM architecture to AI-native products that scale. Engineering intelligence into experiences that matter.
Need this shipped,
not demoed?
Fixed-scope builds, eval audits, or embedded capacity for teams that need agentic AI running in production, not a slide deck. Active DoD Secret clearance for defense-adjacent work.
Production multi-agent workflows and RAG pipelines, designed and shipped end to end.
Find your AI feature's real failure modes, then build the pipeline that catches them.
Embedded senior AI capacity, 2-3 days/week. No six-month ramp-up required.
Trusted by teams at
AI Engineering
at the Principal Level
End-to-end ownership of AI systems. From model selection and architecture through deployment and iteration in production.
Projects That
Ship Intelligence
A mix of AI-native builds and enterprise engineering. Each one solving a real problem with craft.
Eval-as-a-MCP
Built a contact center agent leveraging LLMs and RAG for customer support. Implemented a test and evaluation pipeline to assess model performance and improve response accuracy.
RAG/Document Ingestion Pipeline
Built a real-time document ingestion and retrieval pipeline for AI applications. Implemented a system that ingests documents, processes them into embeddings, and enables efficient retrieval for RAG (Retrieval-Augmented Generation) tasks.
Attest Commerical Real Estate Lease Intelligence
A lease intelligence tool that extracts key terms from commercial lease PDFs, verifies every extracted value against the source document before trusting it, and derives critical dates and risk flags from that verified data — with a citation trail from any field back to the exact page and passage it came from.
Agentic system builds, eval audits, or embedded capacity. Cleared for defense-adjacent work.
What Leaders Say
A track record built on trust, craft, and results, from Fortune 500 to startups.
I had the pleasure of working with Michael at the AWS Generative AI Innovation Center, where he made valuable contributions to AI test and evaluation frameworks and a CI/CD deployment pipeline using GitHub Actions and AWS CodePipeline. His work helped automate the deployment and execution of evaluation processes, making it easier to compare model behavior across versions and validate changes. Michael combines strong technical knowledge with a genuine willingness to help others.He is thoughtful, dedicated and dependable.He consistently made time to collaborate, share what he knew and work through complex technical challenges.Beyond his engineering abilities, he is simply a great person to work with.I recommend Michael without hesitation and would gladly work with him again.
I've worked with Michael for the past several years, first at AstraZeneca and most recently at Amazon, where he was a Software Development Engineer in the AWS Generative AI Innovation Center.He is one of the most knowledgeable AI engineers I've encountered in my career. He is always on the cutting edge of the field and, just as importantly, always eager to help others learn it too.At AstraZeneca, Michael was part of the team that led the introduction of RAG(Retrieval- Augmented Generation) system development to the organization, pairing deep technical expertise with the vision and communication skills needed to bring a new technology into a large enterprise.At Amazon, he applied that same expertise to helping enterprises build and deploy generative AI solutions.On top of all that, Michael is friendly, personable, and a genuine pleasure to work with.I highly recommend him to any organization looking for an engineer at the forefront of AI.
I worked with Michael at AWS, where he quickly got up to speed and made an immediate impact. He is dependable, technically skilled, and always ready to tackle complex challenges. His collaborative attitude and strong problem-solving skills made him a great teammate. He would be a valuable asset to any team, and I’m glad to recommend him.
Michael is a no-drama, smart, and a great problem-solver. Not only did he figure out the system quickly, but his skills proved of great service across multiple aspects of the project.
Michael has very solid experience with deep expertise in modern frameworks. Adept at creating customer-focused, mobile-responsive UI with UX in mind.
Michael is one of the most patient, knowledgeable, flexible and understanding people I have ever worked with. Because of his talents, I still work with him to this day.
Recent Writing
Notes from building agentic AI systems in production. What breaks, what works, and what I'd do differently.
Your eval dashboard is a crime scene photo
Your agent hallucinates a contract clause. Sends it to a customer. Then, an hour later, a dashboard somewhere turns red.
AI Observability Isn't APM With Extra Steps
Uptime and latency can't tell you if an agent's output is correct. Why AI observability needs a different kind of tooling than the APM stack you already have.
Bedrock Agents Classic Just Closed to New Customers. Here's What That Actually Means.
Bedrock Agents is now Classic, closed to new customers, and in maintenance mode — what AgentCore actually changes, and what it means if you're already running agents in production.
Open to New Engagements
Let's build
something intelligent
Whether you're launching an AI product, scaling an existing system, or just exploring what's possible, let's talk.
michaellegemah@gmail.com