Lead AI Engineer · Eindhoven, NL

I build production AI systems people trust and love.

I'm Mykel Schmidt, a hands-on Lead AI Engineer at Flynth. I solve complex enterprise problems with secure, reliable and intuitive solutions—combining AI engineering, Azure architecture, and human-centered product design.

Download résumé LinkedIn Explore my career
Portrait of Mykel Schmidt
ir. Mykel Schmidt MSc Human-Technology Interaction
Current role Lead AI Engineer Flynth · since 05/2026

Career timeline

Scroll down to travel back through time.

  1. Flynth current role · Arnhem / hybrid

    Lead AI Engineer

    Appointed Lead AI Engineer within six months of joining. I stay hands-on while leading the technical direction and development of our AI platform, shaping Flynth's wider AI architecture, standards and governance.

    PythonFastAPIPostgreSQLDocker AzureMicrosoft Foundry
    • Application architecture. Agent/MCP-ready design and UX built to drive adoption, on persistent state, recoverable asynchronous jobs and concurrent workflows behind authenticated, versioned REST APIs.
    • Quality and safety. Structured outputs, grounding + source-linked evidence, PII anonymization, human-in-the-loop design, and more than 1,000 automated tests.
    • Evaluation and operations. (Automated) feedback loops, model benchmarking and A/B testing, with token, cost, cache and timing telemetry.
    • Architecture and governance. Azure Cloud Adoption Framework, DPIA/TIA and ISO 27001 gap analysis, and technical hiring.

  2. Flynth

    AI Engineer

    Joined Flynth to pioneer internal AI applications and build its AI platform.

  3. KeyTakes personal product

    KeyTakes: AI-powered summaries

    KeyTakes

    Product demonstration

    Next.jsRAGSSEFirebase

    Full-stack AI app featuring RAG, LLM integration, and real-time streaming.

    KeyTakes is a full-stack AI application I engineered from the ground up that can summarize any webpage and YouTube video. The system leverages Google's Gemini models to perform complex summarization and analysis tasks. For its interactive Q&A feature, I implemented a Retrieval-Augmented Generation (RAG) pipeline, utilizing vector embeddings and cosine similarity to dynamically fetch relevant context for the LLM.

    To perfect user experience during these long-running AI processes, I architected the backend using asynchronous serverless functions on Vercel and streamed results to the Next.js frontend in real-time with Server-Sent Events (SSE). The project also involved building a robust web scraping service with Puppeteer and managing user data and application state with Firebase.

    Visit KeyTakes.com Summarize myke.works

  4. Eindhoven Airport E-mergo Client

    Data warehouse development

    Contributed to a TimeXtender data warehouse and built reusable patterns for API ingestion, incremental loading and SQL transformations.

  5. Flex IT Rent E-mergo Client

    Multi-agent RAG AI solution

    Led the discovery, architecture and delivery of a multi-agent RAG proof of concept. The pilot received the maximum personal client-evaluation score and approval to continue the initiative.

  6. E-mergo

    Data & AI Engineer

    Joined as a Data & AI Engineer and worked on AI initiatives as well as data warehousing, API ingestion, and BI.

  7. Dawn personal product

    Dawn: New Tab Dashboard

    Wholesome project combining AI & UX, with HTML, TailwindCSS, and JS.

    I created a browser extension called Dawn with the goal of promoting digital wellness. Dawn replaces the standard new tab page with a customizable dashboard where users can select and position cards to their liking. Dawn includes various cards centered around productivity and mindfulness and integrates a Large Language Model (LLM) to enhance its functionality.

    This project allowed me to apply my UX, design, and coding skills, primarily using HTML, TailwindCSS, and JavaScript. It further enhanced my skills in working with APIs and asynchronous programming and offered me an invaluable holistic experience of overseeing an entire development cycle. Dawn received a "Featured" badge from Google representing its adherence to high standards, and was well received by users, reaching 10,000 downloads on the Chrome Web Store.

    Chrome Web Store Try on Web
    Dawn dashboard with its default card layout Dawn dashboard showing the habit-forming cards Dawn dashboard showing an alternative customized setup
    JavaScriptTailwind CSSUXLLM API

  8. Eindhoven University of Technology graduation project

    Master's thesis

    Proposed framework

    AI confidence sets the deferral time

    • 95% sure 9s
    • 75% sure 13s
    • 55% sure 17s

    time = 8 + (100 − confidence) ÷ 5

    Experiment designJavaScriptStatistical analysis

    Confidence-based deferral time in joint human-AI decision-making

    In a deferral system the AI makes an initial decision and hands it to a human for the final call. I proposed using the model's own confidence to set how much time that human gets: less time when the AI is certain, more when it is not, so the decision window itself communicates uncertainty without putting a single number on screen.

    I built the experiment as a web app in HTML, CSS and JavaScript — 40 timed trials with full instrumentation — and ran it on Prolific with 266 participants across four conditions. Showing a confidence percentage made people significantly slower without making them more accurate, and my own framework produced no significant gain either: a null result I reported straight, with what it means for designing uncertainty into AI systems.

    Read the thesis (TU/e repository)

  9. Eindhoven University of Technology Master's degree

    Human-Technology Interaction

    I excelled in 80% of my Master's courses, achieving grades of 8 or higher, while specializing in Human-Technology Interaction. My thesis focused on improving human-AI decision-making by proposing a novel framework that suggests the use of a confidence-based deferral time. I gained hands-on experience in programming and machine learning to design AI models that enhance human-AI collaboration. Key courses like Advanced Cognitive Engineering and Human-Robot Interaction strengthened my understanding of emerging technologies like AI and robotics, setting a strong foundation for my future as an AI engineer.

    Some courses

    • Management of Product Development1ZM16
    • UX Design0HM110
    • Advanced Cognitive Engineering0HM150
    • Human-Robot Interaction0HM280
    • Embodying Intelligent Behavior in Social ContextDBM140

  10. Eindhoven University of Technology Bachelor's degree

    Psychology & Technology – ICT

    My passion for gaming and computers led me to pursue the ICT track in my Bachelor's in Psychology & Technology. This program allowed me to explore the intersection of human behavior and tech, with courses like User Experience for Intelligent Systems and Quantitative Research honing my skills in data analytics, human-computer interaction, and user behavior analysis. I gained solid tech skills through courses in programming, data modeling, and system design, building a strong foundation in both the technical and theoretical aspects of technology. My studies equipped me to design user-focused systems and apply these skills in AI and tech fields.

    Some courses

    • Datamodelling and databases2ID50
    • Logic and Set Theory2IT60
    • Advanced Research Methods0HV110
    • Programming2IP90
    • User Experience for Intelligent Systems0HSUC0

Skills & technologies

My affinity for technology drives my quick adoption of new software, languages, and frameworks—a process I genuinely enjoy.

AI product engineering

Production LLM applications, orchestrated workflows, RAG, structured outputs, human feedback, evaluation and cost/latency control.

Machine learning & data

Classical machine learning next to LLMs—model development, evaluation and deployment—plus statistical analysis, experiment design, data warehousing and BI reporting.

Software & cloud architecture

Asynchronous Python services, APIs, Azure, containers, infrastructure as code, CI/CD, enterprise integration and security by design.

Reliable & responsible AI

Source grounding, deterministic controls, automated testing, model benchmarking, privacy by design, governance and human oversight.

Technical leadership

Architecture and product decisions, engineering standards, cross-functional alignment, technical hiring and business translation—while remaining hands-on.

Human-centered design

An interdisciplinary AI, psychology, UX and research background focused on making complex systems understandable and useful.

Technologies I work with

From the AI stack I build on today, back to the languages, data platforms and research tools I picked up along the way.

  • Microsoft Foundry
  • Microsoft AutoGen
  • LangChain
  • LLMOps
  • Machine learning
  • Agentic AI
  • RAG
  • Structured outputs
  • Evaluation
  • Python
  • FastAPI
  • Pydantic
  • NiceGUI
  • REST APIs
  • JavaScript
  • Node.js
  • React
  • Next.js
  • HTML
  • CSS
  • Tailwind CSS
  • Bootstrap
  • PHP
  • Java
  • Git
  • Microsoft Azure
  • Azure DevOps
  • Docker
  • Bicep
  • Entra ID
  • Azure Storage
  • Firebase
  • Vercel
  • SQL
  • PostgreSQL
  • SQL Server
  • Power BI
  • TimeXtender
  • Qlik
  • R
  • SPSS & Stata
  • Figma
  • SAP Signavio
  • Trello
  • Postman
Dutch Native English Native German Basic

Microsoft certifications

Six certifications earned across AI, data engineering, cloud and analytics.

Azure AI Engineer Associate

AI-102 · Microsoft Certified

Demonstrates the ability to design and implement AI solutions using Azure AI services, Azure AI Search, and Azure OpenAI to meet business needs.

AI

Azure AI Fundamentals

AI-900 · Microsoft Certified

Demonstrates fundamental AI concepts related to the development of software and services of Microsoft Azure to create AI solutions.

AI

Azure Data Engineer Associate

DP-203 · Microsoft Certified

Demonstrates understanding of common data engineering tasks to implement and manage data engineering workloads on Microsoft Azure, using a number of Azure services.

Data

Azure Data Fundamentals

DP-900 · Microsoft Certified

Demonstrates foundational knowledge of core data concepts related to Microsoft Azure data services.

Data

Azure Fundamentals

AZ-900 · Microsoft Certified

Demonstrates foundational knowledge of cloud concepts, core Azure services, plus Azure management and governance features and tools.

Cloud

Power BI Data Analyst Associate

PL-300 · Microsoft Certified

Demonstrates methods and best practices that align with business and technical requirements for modeling, visualizing, and analyzing data with Microsoft Power BI.

Analytics

Get in touch

I'm happy at Flynth and not looking to move, but I'm open to conversations about exceptional opportunities.

Email LinkedIn GitHub View résumé
Eindhoven, Netherlands mjaschmidtfw@gmail.com