
Services
What I've done for years: from first commit to the latest patch.
Build it, make it intelligent, protect it. Same rules every time: written scope, approved changes, rollback, evidence.
Development
Product and platform engineering
Backends, APIs and workflows in PHP and Symfony, Python, Node and TypeScript, or Go. Built to run in production, not only to ship.
Learn more→Integrations
100+ third-party APIs: payments, invoicing, CRMs, e-invoicing. Webhooks, retries and reconciliation that hold up.
Learn more→Infrastructure and DevOps
Docker and Kubernetes, CI/CD, monitoring, cost optimisation and high availability.
Learn more→Legacy modernisation and technical leadership
Upgrades, migrations and stabilised production. Fractional CTO or tech lead, release gates, code review, mentoring.
Learn more→Examples: a service that confirms an airline booking at checkout in seconds, and a custom image service that cut an online shop's image bill by more than 99%. Both are told on the Obsentum page.
AI engineering
Agentic workflows and RAG in production
Assistants and workflows that answer real users, and hand over to a person when they should.
Learn more→Evaluation and observability
Prompt and model changes measured, not guessed. Traces of every run, kept where you can query them.
Learn more→AI-assisted engineering
Agents, skills and hooks with human review as the daily way of building; local models where data must stay in-house, such as the incident triage in fleet-watch.
Learn more→Security and compliance
Breach prevention and hardening
Close the doors an attacker tries first. Penetration testing, attack-surface review and staged patch rollout included. Every change approved, logged, reversible.
Learn more→Social engineering assessments
Signed-off tests of the human process around access. You get fixes, not blame.
Learn more→Compliance readiness
ISO 27001, SOC 2, NIS2, GDPR: owners and evidence that keep pace with new CVEs.
Learn more→The story behind the security work is on the Security page. The monitoring and patching tool is fleet-watch, run for you as a managed service.Managed serviceYou do not install or operate the tool yourself. It is set up and run for you, and you receive the results: alerts, approved updates and records.Learn more →
How an engagement starts
A short conversation
You describe the systems, the worry and the constraints. I tell you honestly whether I am the right person.
A written scope
Targets, methods, time windows, contacts and stop conditions, agreed in writing before work starts.
The work, with checkpoints
Changes only with your approval, each with its rollback.RollbackThe prepared way to undo a change and return to the previous working state. A change without a rollback is a bet.Learn more → You see progress as it happens.
Handover
A report, an evidence trail and a debrief. Retests are included where they apply.
Not sure which service fits?
Describe the problem; I will suggest the smallest engagement that answers it.
Get in touch