Infrastructure engineered for
uptime, security and cost efficiency
I’m Imonikhe. I build secure infrastructure for high-stakes systems.
17+ years across emergency services, energy operations and open source.
Built for the real world.
Reliability at SAHAB. Identity at Greenstand. AI evaluation through Snorkel AI. Open-source platform engineering.
Reliability when every second counts.
Reliability engineering for Saudi Arabia’s digital ambulance platform. Transport requests, emergency coverage, real-time tracking and online payment.
Testing the engineers made of code.
Task authoring and review across Senior SWE-Bench and Terminal-Bench 2.0, 2.1 and 3.0. Testing correctness, reliability and engineering judgment.
Auditable identity. Familiar authentication.
A Keycloak–Fabric identity bridge and Kubernetes delivery for Greenstand’s reforestation platform. Secure access behind a global view of tree captures.
Infrastructure with the operating manual.
Reusable platform packages with architecture notes, recovery runbooks and clear validation boundaries.
Four benchmarks. Different engineering challenges.
Runtime investigation, realistic feature work and codebase conventions.
Explore benchmark Terminal-Bench 2.089 curated tasks with reference solutions and final-state verification.
Explore benchmark Terminal-Bench 2.1Repairs to 28 tasks and continuous validation of the benchmark.
Explore benchmark Terminal-Bench 3.0A harder task set built through open contribution and adversarial review.
Explore benchmarkResearch behind the systems.
Exploring resilience, secure identity and the ways we evaluate intelligent software.
Resilient Kubernetes platforms
Exploring recovery, observability and the operating assumptions behind reliable infrastructure.
Explore the research Security & identityTrust across distributed systems
Examining how authentication and auditable identity connect through Keycloak and Hyperledger Fabric.
Explore the research Frontier AIEvaluating coding agents
Investigating failure modes, independent verifiers and what it takes to recognize a working solution.
Explore the researchBuild for the person
who gets the alert.
I started with servers and satellite links in Warri - Delta State. Working on ambulance dispatch later reinforced the same lesson: infrastructure needs to be understandable when something goes wrong.
More about how I work