research engineer · systems + ai infrastructurelocation: lagos, ng · remotestack: go · typescript · postgres · kubernetesopen source: active — agent-railsfocus: memory · authorization · observabilityresearch engineer · systems + ai infrastructurelocation: lagos, ng · remotestack: go · typescript · postgres · kubernetesopen source: active — agent-railsfocus: memory · authorization · observability

Writing

Technical writing on agentic infrastructure.

Articles, short research notes, and talks. The goal is graduate-level reasoning, not marketing.

Design paper · working draft

Toward Trustworthy Infrastructure for Autonomous AI Systems

The long-form argument behind every project here: memory substrates with provenance, capability-scoped authorization, and observability indexed by decision rather than span.

Read the paper

Articles

2026-07-14·12 min

Designing Memory for Long-Lived Agents

Why retrieval-augmented generation is not enough, and what a real memory substrate for autonomous agents might look like.

Read

2026-06-22·10 min

Why Agent Observability Is Hard

Metrics, logs, and traces were built for human operators. Autonomous agents demand a new observability primitive: the reasoning trace.

Read

2026-05-30·11 min

Runtime Governance for AI Agents

Static allow-lists will not keep autonomous agents safe. Governance must be dynamic, uncertainty-aware, and grounded in simulated impact.

Read

2026-04-18·14 min

Building Reliable Multi-Agent Systems

Coordination, conflict resolution, and shared context are the defining challenges of multi-agent systems.

Read

2026-03-12·9 min

Infrastructure Lessons from Production AI

What running large-scale ML systems taught me about building agentic infrastructure.

Read

Research Notes

2026-08-02·4 min

Thoughts after reading the MCP specification

MCP is a promising step toward standardized tool use, but it leaves hard questions about state, identity, and trust unanswered.

Read

2026-07-28·3 min

Memory shouldn't be treated as context

Stuffing history into a prompt is a hack, not an architecture.

Read

2026-07-10·5 min

How agent evaluation should evolve

Benchmarks that measure single-turn correctness miss the point. We need to evaluate agents across episodes, not just turns.

Read

Talks

2026-08-15·45 min

Toward Reliable Infrastructure for Agentic AI

A systems-level view of the infrastructure needed for trustworthy autonomous agents.

Watch