Working software in production, agents that do real delivery work, and technical direction from someone who still builds.
I build AI-native, agentic product and software systems end to end and stay hands-on, from technical direction and architecture to delivery and operations.
Start a conversation · Projects and CV
- A product that reaches production and keeps working. Thin end-to-end slices, tests first, and every release checked where it runs.
- Agents that run the delivery loop, not just write code. They carry work from signal through build, review, merge, deploy, and the check in production; people set direction and investment boundaries.
- Decisions grounded in the running system. Product judgment, architecture, and technical direction from someone who builds, ships, and runs the software.
- A system your team can own. Clear ownership, readable code, decisions written down, observability, and backups with drilled restores.
- CTO-level scope, proven. As CTO I scaled remote delivery across 15 teams and 100+ developers, and as hands-on CTO I built several products from 0 to 1.
- Still building, at speed. My own product Jev Studio went from first commit to live in one week, built and run by agents.
- Agentic engineering in the open. My Skills & Harness runs across Claude Code, Codex, Cursor, and Gemini CLI, and the levels I work by are public as the Agentic Engineering Maturity Model.
- My own system already runs this way. Across a dozen repositories, from products and a public website to a homelab platform, agents pick up work, build it test-first, get it reviewed in fresh context, merge, deploy, check it in production, and write back what they learned. They also keep dependencies and their own Skills current and drill restores; people step in for direction, budgets, and anything that leaves the system.
- Taking it beyond delivery. I am extending the same loop to the whole product: agents that turn signals into triaged problems, prepare decisions, ship, and review the outcome, so the next step follows the evidence.
- Start from the problem and the value it is worth, then build the smallest end-to-end slice that proves it.
- Every practice earns its place by improving flow, quality, or decisions; anything else goes.
- Quality, security, design, and operations join the first conversation, not a late gate.
- Done means running in production and checked there.
- Autonomy grows as fast as the evidence does: tests first, independent review, required checks, and a stop path for every loop.
Engineering Director, Agentic AI at OH-SO Digital. I work where client problems, product ideas, and technical reality meet: technical responsibility across client projects, hands-on, and helping shape how OH-SO builds software and shares what it learns.
- Agentic Engineering Skills & Harness - the public Skills, repository patterns, review loops, and product-engineering model I use to make agent work reliable, from bounded repository tasks to governed value streams.
- Jev Studio - decision rules for incoming text, written as plain questions, checked on labeled cases, and handed to engineering as typed code with tests. (project page)
- As CTO, scaled remote delivery across 15 teams and 100+ developers, using flow signals, golden paths, cloud-native delivery, and stronger operational ownership
- As hands-on CTO of an AI-supported geodata product platform, owned product and UI work, worker flows, geo and routing services, product research with evidence gates, and platform operations
- As hands-on CTO and Chief Product Owner, built an AI-supported freelancer-project matching platform from 0 to 1, with product, technology, and team setup in one responsibility
- As fractional CTO after product-market fit, advised on technology selection, team building, architecture, operations readiness, and scaling direction
- As hands-on CTO, built a health-study SaaS from product model through backend, frontend, CI/CD, team setup, and operations
- Built SERP parser and crawler systems in a SaaS SEO analytics context, including parser work for 450M+ keywords worldwide
- Led backend, delivery, and architecture work for a big-data SEO platform with Java/Scala, microservices, multi-database architecture, AWS/Docker platform work, and team leadership
- Languages: TypeScript, JavaScript, Java, Kotlin, Python, Scala
- Web and product: Astro, Nuxt, Vue, React, Svelte, Next.js, Tailwind CSS, PWA patterns, REST/OpenAPI
- Backend and data: Spring Boot, Play Framework, FastAPI, Bun, BullMQ, PostgreSQL, PostGIS, MySQL, ArangoDB, MongoDB, Redis
- Delivery and operations: CI/CD, GitHub and Gitea Actions, Docker, Kubernetes, Traefik, Renovate, observability, alerting, backups with drilled restores
- Agentic and AI work: Claude Code, Codex, Cursor, typed decision models, evaluation and evidence gates, workflow automation, NLP
If you are building agentic product or delivery systems, or need hands-on technical direction for a product that has to work in production, start the conversation on daume.dev.