Richard Okonicha

Architected sovereign AI cloud platform on bare-metal OpenStack with GPU/TPU compute fabric; engineered LLM gateway with custom
model-routing, LiteLLM, and Helicone analytics. Built orchestration layer with SkyPilot, automated failover, and browser-based terminal
(xterm.js). Reduced inference latency from 300ms+ (EU) to <50ms (Lagos); cut tenant operational costs by 65% via microservices
migration

- Build and maintain core AI integrations enabling developers to query LLMs via SQL workflows. Collaborate with the core engineering team
on production-grade model serving and automated fine-tuning pipelines.
Stack: Python, SQL, MindsDB Core, Model Serving, LLM Integrations

- Architected AI adoption strategies for enterprise transition to agent-driven systems. Implemented Turbo Quant quantization on self-hosted
LLMs, optimizing VRAM and inference throughput. Built production AI pipelines with LoRA fine-tuning, function-calling accuracy
optimization, and agentic harnesses with guardrails and human-in-the-loop fallback.
Stack: Python, Turbo Quant, LoRA, OpenClaw, Herme

- Led enterprise application architecture for Swiss AI writing platform with strict multi-tenant isolation. Resolved high-concurrency race
condition in async request queue, eliminating production 500 errors. Redesigned request pipeline for enterprise traffic spikes; served as
technical liaison between stakeholders and engineering teams.
Stack: Python, Tenant Isolation, High-Concurrency Queues, Doc

- Owned the core storefront theme engine and spearheaded feature development for a next-generation, high-scale enterprise SaaS platform.
Designed a resilient Stripe payment architecture processing thousands of concurrent global transactions daily. Refactored the checkout
pipeline to eliminate distributed state race conditions, resolving duplicate transaction charge bugs. Built a containerized clie

- Designed production ML pipelines forecasting software delivery timelines; built predictive analytics engine reducing forecast error by 30%.
- Architected full-stack interfaces and integrated ML engine into client dashboards via high-performance REST APIs.
Stack: Python, React.js, Docker, PostgreSQL, REST APIs, Feature Engineering

- Built high-throughput REST API backend for large-scale contact-tracing/exposure notification application (By Linux Foundation in partnership with Kiel university; COVID19 ). Isolated and
patched a critical memory leak in the upstream MongoDB driver causing system degradation under peak traffic loads. Optimized database
query paths and connection pooling to ensure uninterrupted availability durin

Led team of 6 engineers building health data systems for government clients. Designed OpenHisa health insurance system deployed
across 12 Nigerian states; engineered MSDAT platform (Gates Foundation) serving 10,000+ users. Refactored legacy database paths,eliminating N+1 query patterns to reduce critical API latencies by 40%.
Stack: Python, Django, React.js, MongoDB, PostgreSQL, Docker, Kubernetes

- Selected for international accelerator focused on agile development and VC fundraising. Architected and shipped multiple MVPs from
scratch in Python and JavaScript under constrained timelines.
Stack: Python, JavaScript, Shell Scripting, MVP Architecture

Advanced Computer Science Concepts
Memory Management
Algorithms
Data Structures
Arrays
Database Systems
Problem-Solving Techniques
Web Technologies
Programming Languages: C and Python
Cloud Fundamentals
Distributed Systems
Course Director: Dr. David Mulan (Ph.D.)

Startup Involvements:
Roar Start-Up Incubation Hub - Cohort 4
Rotary Club
Cedars Group
Nigerian Universities
Startup Creation
Best Time Ever!