I build
Agents you can
trust to act.
Agent systems end to end — and the governance and guardrails that make them safe to put into production.
What agent problem are you facing?
Describe it — this live agent scopes how I'd build it.
proprietor · Circle AI Agents Hackathon Berlin (1st)
An AI agent that owns and operates a profitable micro-SaaS company.
It sells a Company Enrichment API, fulfilled by a multi-agent research pipeline: parallel domain researchers under a central orchestrator, entity disambiguation before any expensive call, iterative knowledge-gap loops. The business runs itself around it: customers pay retail through an HTTP 402 paywall, and revenue and costs settle through an Agent Wallet.
The hard part wasn't proving an agent can run a business — it was making one you'd trust to leave alone: a spend gate on every payment (daily cap, runway floor, fail-closed approval above a threshold), decline-before-charge on unprofitable work, and a per-job P&L ledger recording every decision on-chain.
agentplane · AI Workshop Berlin
Self-hostable, framework-agnostic governance layer that sits inline between AI agents and company systems — enforcement, not after-the-fact observation.
Every tool call is checked against a policy engine before it runs. When a call crosses a line, the gateway holds the connection open and waits for a human. A context server exposes the company's own rules and precedents to every connected agent, so decisions are made against institutional knowledge rather than a generic model's guess. Every decision is written to an audit trail that captures the full context — inputs, matched rule, grants evaluated, and who resolved it.
agent_workspace · QESTIT
Internal agent IDE for authoring and executing custom agent skills.
Skill system with runtime discovery, SSE streaming for long-running agent operations with mid-session question/permission handling, context integration (Jira, Confluence). Three-layer evaluation framework: execution trace parser with subagent timing and parallelism metrics; deterministic quality calculator across coverage, traceability, syntax validity, reusability; parallel LLM judges with externalized rubric prompts and contrastive integration tests.
agent_analytics · QESTIT
Per-call token, cost, and latency tracking across every agent LLM call — the observability layer for what an agent spends.
Request analytics service streaming daily cost trends, per-model breakdowns, and usage comparison across all AI API calls. Token-counting service built from scratch: tiktoken with the OpenAI Vision token formula (image scaling, tile computation, detail modes), pre-request estimates validated against actual API usage.
browser_automation_agent
Browser agent that navigates and completes forms on sites it has never seen — LLM vision instead of per-site selectors.
8-state navigation FSM, two-stage field resolution combining fuzzy semantic matching with LLM content generation, multimodal screenshot analysis via Llama-4-Maverick. Tracks per-step success/failure to prevent navigation loops. Runs across arbitrary company career sites with no platform-specific code — the failure modes are the point: unseen layouts, cookie walls, dead ends.
task_automation_system
Hybrid RAG end to end — dense + sparse retrieval, query expansion, time-decay reranking, and a fine-tuned T5 extraction model in production.
Weaviate hybrid search (BM25 + dense) with configurable alpha, query expansion via synonym generation, post-retrieval reranking with time-decay × priority-boost. T5 fine-tuned on synthetic task scenarios, 4-bit quantized, deployed on Cloud Run with GPU.
Most AI agents work in the demo and fail in production. They pass the happy path, then break quietly — a tool call returns garbage and nothing notices, a prompt tweak regresses something nobody was measuring. That gap is what I work on: evals, observability, and the engineering that makes an agent trustworthy. State machines instead of vibes, eval frameworks instead of "looks good to me."
I'm CTO of Compass, building through Founder School 2026 and Emprelatam. Compass is AI agent security for the moment agents start moving real money — the risk is simple: one manipulated input, and an agent pays out in a way its owner never intended. I build the layer that stops that, deciding what an agent is allowed to do before money moves, and proving what it actually did after. The design assumes manipulation: even the LLM judge in the decision path can only fail toward deny, never toward allow.
Previously as a Senior AI Consultant at QESTIT I built the orchestration and evaluation tooling behind a team of in-house agents and coached engineers on adopting LLM workflows — the full arc from "we should try agents" to systems a business runs on every day.
Before agents I was a blockchain engineer, working on production systems where mistakes are irreversible. At Civic Technologies I shipped the first production app built on Internet Identity's then-undocumented blockchain identity stack — a secure, private credential-sharing system for identity verification (Rust, TypeScript and Node). At Parity Technologies I wrote the technical documentation for Polkadot's cross-chain messaging system and its Rust developer tools, and was selected for the Polkadot Academy with Gavin Wood (co-founder of Ethereum). At Sunrise Stake I wrote high-performance smart contracts for cross-chain transactions used by Solana to offset the chain's carbon emissions.
1st overall prize at ETHGlobal Paris 23 (1000 hackers), first prize at ETHGlobal New York 23, ETHPrague 23, ETHGlobal Istanbul 23 (×2), and NFTBerlin 22 (×3). Graduated with my Computer Science BSc degree in October 2023.