Applied AI & Agentic Systems Engineer | Backend & Systems Developer
Akure, Nigeria • LinkedIn • Email

Dual-degree engineer specializing in autonomous agent architectures, model context protocol (MCP) tooling, compiler-driven protocol generation, and low-latency systems. Experienced in architecting production-grade workflows that bridge frontier multimodal LLMs with deterministic backend infrastructure and adversarial security constraints.

• B.Tech in Cybersecurity — Federal University of Technology Akure (FUTA) • CGPA: 4.13 / 5.00
• B.Sc in Computer Science (First Class Honours) — University of the People (UoPeople) • CGPA: 3.56 / 4.00

• Languages & Core: TypeScript, Python, Node.js, Go, Solidity, SQL, C/C++
• AI & Agentic Systems: Model Context Protocol (MCP), Gemini Multimodal (Vision & Spatial Reasoning), Tool Calling & Function Orchestration, Atlassian Rovo, Adversarial Multi-Agent Governance
• Systems, Networking & Performance: Ultra-low-latency streaming, Socket protocols, Parser/Compiler AST generation, Property-Based Testing (Fast-Check), Asyncio concurrency
• Cloud & Infrastructure: Docker, GitHub Actions CI/CD, Azure Web Services, Linux/Ubuntu systems, SQLite, Redis, PostgreSQL
• Security & Verification: Threat modeling, Zero-Trust architecture, ERC-4337 Account Abstraction, Groth16 ZK-proof verification

Next-generation meta-programming system and protocol compiler

• Compiles declarative YAML protocol specifications into type-safe parsers, serializers, clients, and servers across 4 targets (TypeScript, Python, Go, Rust).
• Implements semantic AST optimization and common-prefix state-machine merging.
• Generates native Model Context Protocol (MCP) servers to expose binary and custom wire protocols directly to AI agents.
• Verified with property-based testing (Fast-Check) to guarantee serial round-trip integrity (parse(serialize(x)) == x).

Ultra-low-latency, zero-install PC-to-mobile remote controller

• Native Windows screen capture and streaming pipeline with adaptive frame throttling and sub-50ms latency.
• Full touch-to-trackpad gesture translation and bidirectional input synchronization.
• Zero external client installs required; operates entirely through modern browser endpoints.

Verifiable AI Treasury Agent for MNEE Stablecoin

• Autonomous B2B financial orchestration on Base L2 using ERC-4337 Account Abstraction for gasless transactions.
• Dual-agent adversarial governance loop (Red Team vs Blue Team AI Auditor) with Groth16 ZK-proof generation for private invoices.
• Native conversational interface integrated directly across Telegram and Slack.

Multimodal AI flood disaster response command center

• Uses Gemini 3 Flash to estimate real-time flood depth from surveillance and drone footage using physical reference anchors (wheel hubs, structural markers) to eliminate physics hallucinations.
• Dynamic supply-chain rerouting engine (e.g. switching vehicles from trucks to canoes based on water depth).
• Hierarchical Monitor-Predict-Plan-Execute-Verify decision loop with SHA-256 state hashing.

Telemetry machine learning platform for motorsport performance

• 6-model predictive ensemble (XGBoost, CatBoost, LightGBM, Random Forest) for lap-time forecasting with 0.51% MAE (0.513s).
• Evaluates 500+ telemetry records across speed, throttle, brake, and lateral G-force to generate driver sector insights.

Atlassian Forge enterprise parts lifecycle & cost-cap monitor

• Built for Williams Racing during Atlassian Codegeist.
• Integrated Rovo AI Agent for natural-language telemetry querying and inventory readiness scoring.

• Deterministic Verification: LLM reasoning must always be bounded by typed schemas, test assertions, and strict human/agentic gates.
• Low Latency & High Signal: Clean, minimal abstractions over raw sockets and protocols; zero bloat.
• Security by Design: Threat containment, principle of least privilege, and auditable transaction trails across all autonomous pipelines.