Segun Olumbe

Background

Day job at NBCUniversal, toolkit, and the engineering behind the tools.

  1. Day job

    nbcuniversal · ott

    Part of the team delivering streaming services to millions of consumers across Peacock, NOW, and SkyShowtime at NBCUniversal. I implement and optimise OTT clients for TVs and consoles (PlayStation, Xbox) in high-traffic environments. Recently lead developer on the Telemundo dual-language project, bringing a Spanish experience to Peacock for the 2026 World Cup. I sit at the intersection of technical implementation and user-centric design: clean, maintainable code, informed by design theory so interfaces stay intuitive on a remote.

  2. Stack

    whatever fits

    I am not bound to one language. I use whatever solves the problem and learn what I do not know yet. Day to day that is often LightningJS, React, Redux, TypeScript, GraphQL, Node.js, Bash, and Python. Pipelines and runtime: Kubernetes, Docker, Jenkins, GitHub Actions, and Concourse CI. Outside work I also run things on VPS hosts and AWS, plus Cloudflare, Stripe, and Paystack when a side project needs them.

  3. Compliance

    controls that ship

    I like compliance when it is concrete: versioned rules you can diff, checks that fail in CI, evidence you can hand to someone. Spare-time work includes bouncer for UK Online Safety Act and ICO Children's Code controls, and aiglare for AI governance mappings (SOC 2, EU AI Act, NIST AI RMF, OWASP LLM Top 10). I would rather encode a control than argue about it in a slide deck.

  4. AI

    generation + trust

    Frontier models already plan, search, edit, and recover. The durable work is the independent trust layer around them: deterministic context before the edit, exact validation after it, permissions, review evidence, and a human merge decision. Solumbe is that trust harness; gate, tieline, bouncer, and aiglare extend the same fail-closed approach across delivery, contracts, compliance, and AI governance. On the product side, BashBop's event assistant answers from retrieved records, with a deterministic fact check and evals. I have also built agents on the Anthropic SDK in three patterns: a ReAct loop, an orchestrator with specialists, and generate-then-evaluate.