David A. Schirmer
Los Angeles, CA | dschirmer@komatik.ai | linkedin.com/in/david-a-schirmer | komatik.ai | david.komatik.ai
Founder & CEO, Komatik | AI Product and Platform Leader
Summary
Founder and CEO of Komatik, an AI company that builds, hosts and runs custom software with an AI workforce under human oversight. Left Accenture in February 2026 and took the company from first commit to a ten-product portfolio in eight months, owning strategy, product, brand, infrastructure and budget.
A career of widening scope: a decade in the U.S. Navy, rising to First Class Petty Officer and leading air traffic control operations at sea and ashore, then enterprise data science, AI and cloud architecture for top-tier clients at Accenture, and now company leadership and hands-on AI platform engineering at Komatik. Runs the company on one rule: AI agent teams handle the volume, people make the decisions, and every output carries evidence.
Experience
Komatik | Los Angeles, CA | 02/2026 - Present
Founder & CEO | 02/2026 - Present
Founded an AI company that builds, hosts and runs custom software for businesses, with AI agent teams doing the work and people accountable for every decision. As sole founder, owns strategy, product, brand, engineering, infrastructure and budget.
Company leadership
- Took Komatik from first commit to a 10-product portfolio in 8 months: the core platform (AI-built custom software), Trace (AI app readiness checks), Trailhead (release gating), Pack (tool catalog for AI agents), Slipstream (AI context assistant), Lodge (workspace canvas), Sundog (AI game development), Firn (AI narrative platform), Lyra (AI music streaming) and Cairn (open agent research).
- Built the business from zero: defined the business model, offer and go-to-market for an AI-run software company, and stood up the full customer path from sign-up through live billing to delivered software.
- Designed and ran an AI workforce instead of an early hiring plan: 21 role-based agents (coordinator, security QA with veto power, release manager, payments) with risk levels, per-agent budgets, access controls and human approval gates.
- Owned infrastructure and spend across a multi-cloud stack spanning hosting, databases, edge compute, payments and AI model providers, and built internal AI workloads on NVIDIA Blackwell architecture.
- Set the governance bar: a 10-area security audit, a CI security gate that blocks critical findings, database access-policy tests, staged dev to staging to production releases and a weekly backup-restore drill.
Product and engineering
- Architected the platform end to end: a business orders custom software, AI agents plan and build it, and Komatik hosts, monitors and maintains it on a monthly plan, with each client's data kept separate.
- Built the engine that turns an order into a running app: it plans the work, has AI agents assemble it in any of 20 framework families, from web and mobile apps to back-end services and device firmware, then tests it in isolation and deploys it.
- Built the AI team layer: each customer gets a team of specialist agents (analyst, developer, marketing, payments, infrastructure) that take on 30 kinds of business work from a shared queue, coordinate with each other and hold finished work for the owner's approval.
- Built an AI coding agent that turns a customer's request into a plan, makes the code changes, then checks and repairs its own work before handing it back.
- Built model routing that sends each task to the AI model that does it best, trials new models on a small share of work before rolling them out, and tracks how often work is right the first time and what it costs.
- Shipped Trailhead, an open-source tool that decides whether a release is safe to ship from its test results, risk and live production health, now in its fourth version.
- Launched four consumer products: Lodge (beta, web and Windows), Sundog (beta, Windows, macOS, Linux), Firn (early access) and Lyra (live).
- Built Vector, an AI air traffic control simulation covering 6 FAA en route centers and 22 sectors on live aircraft position data, drawing on Navy air traffic control experience.
- Built the company with AI coding agents (Claude Code, Cursor) writing about three quarters of the code, every change under founder review.
Accenture | Los Angeles, CA | 09/2023 - 02/2026
Data and AI Consultant | 2025 - 02/2026
Business Architecture Specialist | 2024 - 2025
Technology Specialist | 09/2023 - 2024
Hired through Accenture's veteran recruiting and promoted twice, to Business Architecture Specialist and then Data and AI Consultant, while moving from cloud work at an energy utility to retail e-commerce to leading data science for a global payments business.
- Lead data scientist, then data and AI lead, for the marketing arm of a large web search provider's payments business for a year. Ran payment analytics across the whole company, covering every user and every payment action in every country, tracked to the millisecond across billions of data points, with user profiling and segmentation that shaped the product around the features people need and use.
- Designed a dynamic communication layer across email, in-app and text, with messages built for each customer segment and its daily usage.
- Worked across the company's internal teams with frontier LLMs. Redesigned the accounting and metrics workflows behind millions of daily users, and brought the company's frontier AI models into the management and reporting layer to speed up analysis and international strategy across the Americas, the UK, Japan and China.
- Worked on site at a large specialty clothing retailer's headquarters with a small team to modernize and redesign its website and checkout in 2024. Online sales went from down 9.2% over the first three quarters of the fiscal year to up 6.4% in the fourth.
- As an additional assignment between the utility and the retailer, designed a 24/7 IT support model for a major airline using offshore teams, with a staffing plan that lowered operating cost and improved reliability.
- Automated a Southern California energy utility's trouble-call system across California on Azure. Led on Azure security and worked directly with the cloud provider's engineers on a platform where residents and employees report, triage and fix issues of any size across the power grid.
- Architected a multi-cloud database prototype to move the utility's legacy on-premises systems to AWS, Azure and GCP.
United States Navy | Point Mugu, CA | 09/2011 - 05/2021
Air Traffic Controller First Class (E-6) and Facility Watch Officer | 09/2011 - 05/2021
Ten years from special boat training in Coronado to First Class Petty Officer running air traffic control at sea and ashore: deployed with Tactical Air Control Squadron 22 on USS Iwo Jima, then ran the watch at Point Mugu.
- As facility watch officer at Point Mugu, ran operations for a $100M naval air traffic control facility through UAV testing and long-range surface-to-sea missile tests off Southern California.
- Deployed aboard USS Iwo Jima with Tactical Air Control Squadron 22 in 2014 during Operation Enduring Freedom, including anti-piracy operations off Yemen and Oman and exercises with Jordanian, Israeli, British and French forces.
- Started in the SWCC pipeline in Coronado; after a knee injury, cross-rated to air traffic control through ATC school in Pensacola and rose to First Class Petty Officer (E-6).
- Set risk-mitigation protocols and operating standards for deployments abroad.
- Navy and Marine Corps Achievement Medals (2), Battle Efficiency Award and Expeditionary Warfare Excellence Award aboard USS Iwo Jima (2014), Sailor of the Year (2017).
Education
- M.S., Data Science (in progress) | University of Southern California
- M.B.A., Business Analytics | Pepperdine University | 2021-2023
- B.S., Aeronautics | Thomas Edison State University
Certifications
- Python for Fintech Applications | Pepperdine University
- Data Science Certification | Divergence Academy
- Veteran Peer Support Advocate | U.S. Department of Veterans Affairs
Skills
AI and agents: Multi-agent orchestration, MCP servers, LLM evaluation, Model routing, RAG (pgvector), Human-in-the-loop design, Local inference, vLLM, llama.cpp
Languages: TypeScript, Python, SQL
Platforms: Next.js, React, Supabase (Postgres), Vercel, Fly.io, AWS, Azure, Google Cloud, Stripe
Engineering: GitHub Actions CI/CD, Vitest, Playwright, pgTAP, Security review, Docker
Analytics: Tableau, Predictive modeling, A/B testing, Customer segmentation