Now

What I'm focused on right now.

• Closed the Agent Development Lifecycle with two new builds, both on this site now: PreFlight, an LLM-free in-org scan that scores whether a customer's real case history can carry an agent before a single build hour (79/79 tests, 93% coverage, measured); and Nabz, a zero-egress runtime watchdog that correlates a live agent's quality drops with the org's change stream and grades the cause A–E (126/126 tests green on the live pilot org). Ten case studies total — and no new projects planned: the work now is making these visible.
• Configra — my Salesforce ISV product (an RLM managed package) — is in AppExchange Security Review. That review is the gate between 'a package I built' and 'a product customers can install'.
• Going deeper on agentic delivery: I designed an agent loop that builds a Salesforce org end to end — six custom MCP deployment tools, ApexTestRunner as the verifier, deploy → test → fix → repeat until green. The 24-minute MCP walkthrough and the Hospital-org case study are both on this site now.
• HanseWatt's Agentforce agent now answers grounded and takes real action — it identifies the customer, explains a Data 360-grounded consumption anomaly, and logs a support Case. Next: identity resolution and LWCs.
• Built the full trust layer around that live agent — all three now on this site with real output: Prüfstand, a pre-registered German red-team harness with a deterministic verifier the model never controls; Agent Blast Radius, a static analyzer of what an agent's code can actually reach, with a live org as the judge (216 tests); and Aktenlage, an EU AI Act Article 26 evidence pack built from an agent's traces — zero LLM, zero credits.
• German: B1 certified, working toward B2 — the level that opens most DACH job descriptions.
• Actively applying: open to Salesforce Developer / Forward Deployed Engineer roles in the EU and DACH region. I would need visa sponsorship — relocation-ready, targeting Cologne / NRW.

Last updated: 2026-07-22