I'm an AI/ML & Automation Engineer and Researcher with a Chemical/Process Engineering background, specializing in intelligent systems, agentic workflows, and applied machine learning. I design LLM-powered pipelines, retrieval-augmented generation (RAG) systems, and AI-driven automations across fintech, renewable energy, insurtech, FMCG, R&D, power, and Oil & Gas. I've shipped impactful solutions for CatalyzU (TechStars '23), BlackCod Group (Bluebulb), Curacel (YC '22), TotalEnergies, Transcorp Power, Nestlé, and NitHub (UNILAG Innovation Hub).
Proficient in Python, JavaScript, SQL, TensorFlow, PyTorch, LangChain, LlamaIndex, Hugging Face, n8n, Make, Zapier, and Power Automate, I build fine-tuned LLMs, MCP and tool-calling pipelines, and end-to-end automations on AWS and GCP. I'm passionate about building reliable, trustworthy, multimodal, and scalable AI systems, and I'm always open to collaborating on projects that push the boundaries of intelligent automation and digital transformation.
AI/ML & Automation Engineer and Researcher - building intelligent systems where AI, engineering, and impact meet.
I'm an AI/ML & Automation Engineer and Researcher who designs and ships LLM-powered pipelines, retrieval-augmented generation (RAG) systems, agentic workflows, and end-to-end automations. My foundation is unusual and deliberate: a degree in Chemical & Process Engineering from the University of Lagos paired with a degree in Applied Artificial Intelligence from IU, Germany - so I understand both the physical processes that power industry and the intelligent systems that optimize them. That work has been backed by four national scholarships and recognition as a Millennium Fellow, Developer's Foundry Fellow (AI/ML), and Student Energy Fellow.
My work spans the full machine-learning lifecycle - AI engineering, ML engineering, MLOps, data science, and automation. I fine-tune and evaluate LLMs (Qwen3, Llama 3.1) with QLoRA/Unsloth, build RAG and MCP tool-calling pipelines with vector search and guardrails, train computer-vision models for real-world safety and detection, and architect n8n and Python automations on AWS and GCP with CI/CD. I've open-sourced 13+ models and datasets on Hugging Face (with 25K+ dataset downloads) and shipped live products on Modal, Streamlit, and HF Spaces.
I've delivered impact across fintech, renewable energy, insurtech, FMCG, power, and Oil & Gas - designing agentic AI workflows at CatalyzU (TechStars '23), automating operations across subsidiaries at BlackCod Group (Bluebulb), building LLM automations at Curacel (YC '22), driving digital transformation at TotalEnergies, and researching applied AI at NitHub (UNILAG Innovation Hub), with earlier process-engineering roles at Transcorp Power and Nestlé. The constant: I use AI, automation, and engineering to turn manual, inefficient operations into reliable, measurable systems.
But technology is only a tool; my real passion lies in multiplying its impact through people. Leadership, for me, is a platform for service and building communities. This conviction has driven me to take on roles such as President of the UNILAG Energy Club, where I led a team to achieve over 600% growth; VP of the Nigerian Society of Chemical Engineers (UNILAG), where I supported over 300 students; and Sponsorship Lead for GDSC UNILAG, securing over ₦5M for student tech projects. Through Project KUJUA, an initiative I founded, we have provided career mentorship to over 10,000 students in rural communities. My commitment to leadership and innovation has been recognized through global programs like the Millennium Fellowship, Student Energy, and the Developer's Foundry Fellowship (AI/ML).
Looking ahead, my vision is to become a business and global leader dedicated to building AI-powered solutions that accelerate the energy transition and support the growth of climate-tech. I want to be at the forefront of creating a more sustainable and equitable future.
I am always eager to connect with fellow innovators, founders, and leaders in the AI, energy, and climate-tech spaces. If you are passionate about building solutions that matter and are looking to collaborate on transformative projects, please feel free to reach out.
"Remember, one can chase a thousand, but with AI, one can chase a million."
B.Sc. Chemical Engineering
University of LagosAI/ML Engineer
CatalyzU (TechStars '23)Founder, Project KUJUA
10,000+ Students ReachedIU International University of Applied Sciences
March 2025Completed degree in Applied Artificial Intelligence, focusing on cutting-edge ML techniques, deep learning architectures, and practical AI applications.
Pursuing a degree in Chemical Engineering with focus on process optimization and integration with AI/ML technologies.
Intensive 12-month fellowship program focused on advanced AI/ML development, practical applications, and real-world project experience.
6-month comprehensive training program on sustainable energy, renewable technologies, and the global energy transition landscape.
Designed and deployed agentic AI workflows using n8n, Google Workspace, Slack, Notion, and Python to automate ways of work, project management, onboarding, content curation, community management, and internal reporting - reducing manual effort and time wasted by ~60%.
Design, build, and deploy end-to-end business automations across multiple subsidiaries, streamlining operations in fintech, HR, customer support, treasury, internal control, risk, and compliance workflows.
Architected and deployed AI agents and intelligent automation systems to streamline business operations and reduce manual workflows in Y Combinator-backed insurtech company.
Supporting Deep Water Operations for Akpo and Egina Assets, applying advanced process engineering, maintenance strategies, and flow assurance principles to optimize offshore oil & gas production.
Architecting and implementing proof-of-concept AI models for tech incubation projects while contributing to AI/ML education and community research initiatives.
Participated in elite 6-month training program with 25 global peers, developing expertise in sustainable energy transitions and conducting impactful research on renewable energy adoption in Africa.
Selected among top 10 students nationwide for elite engineering internship at 972MW gas power plant, focusing on plant efficiency, reliability, and advanced power generation technologies.
Drove process optimization and efficiency improvements in state-of-the-art automated beverage production facility, delivering measurable results in downtime reduction and production efficiency.
Leading AI/ML research and content development, delivering high-impact technical documentation and insights across diverse AI domains.
Drove machine learning research and development projects, delivering high-accuracy predictive models and contributing to AI-driven solutions.
UNILAG Energy Club
Jul 2024 - Present · 1 yr 4 mosLed organizational transformation achieving 800% membership growth and establishing the club as a premier energy transition hub for students.
Nigerian Society of Chemical Engineers (NSChE UNILAG Chapter)
2023 - 2024Supported 300+ chemical engineering students through academic excellence programs, professional development initiatives, and industry networking.
Google Developer Student Club (GDG On Campus UNILAG)
Sept 2023 - Aug 2024 · 1 yearSecured over ₦5 million in sponsorships for student tech events, hackathons, and community programs at UNILAG.
United Nations Academic Impact & MCN
2023Selected as one of Nigeria's Millennium Fellows, a prestigious global leadership development program focused on advancing the UN Sustainable Development Goals.
Project KUJUA
Oct 2022 - Present · 3 yrs 1 moEducational initiative empowering over 10,000 students with knowledge and resources in technology, entrepreneurship, and personal development.
Google Developer Student Club (GDG On Campus UNILAG)
Sept 2022 - Aug 2023 · 1 yearLed program planning and execution for GDSC UNILAG, managing a team of 8 people and organizing major tech events for student community.
Professor Ayodele Awojobi Design Competition (PAADC)
May 2024 - Present · 1 yr 6 mosAchieved the highest application in competition's history through strategic external partnerships and campus engagement across Nigeria.
Shell Eco-Marathon Team UNILAG
Sept 2022 - Dec 2023 · 1 yr 4 mosLed UNILAG's Shell Eco-Marathon team, coordinating technical development, logistics, and competition strategy for energy-efficient vehicle design.
The Energy Club, University of Lagos (UEC)
Feb 2020 - Sept 2021 · 1 yr 8 mosLed research initiatives for UNILAG Energy Club, coordinating energy studies, technical publications, and knowledge dissemination.
Shell Eco-Marathon Team UNILAG
Jun 2021 - Sept 2022 · 1 yr 4 mosManaged all communications, public relations, and social media for UNILAG's Shell Eco-Marathon team.
Nigerian Society of Chemical Engineers (NSChE UNILAG Chapter)
Jul 2021 - Jul 2023 · 2 yrs 1 moManaged student welfare programs and support systems for UNILAG chemical engineering students over 2 years.
AI/ML Communities
2021 - PresentModerating and fostering engagement in AI/ML communities, facilitating knowledge sharing and collaboration.
Various Initiatives
2019 - PresentContributing to educational and social impact initiatives, promoting STEM education and youth development.
Open-source models, AI/ML applications, automations, and data-science work - filter to explore. Every item links to a live demo, model, or repository.
Nigeria's first domain-specific LLM for Oil & Gas. Qwen3 fine-tuned with QLoRA/Unsloth on 33,000+ instruction pairs, paired with a self-updating dynamic RAG assistant. Served on Modal (vLLM) and Hugging Face Spaces.
User-modelling & recommendation agent for review simulation (InsideNaija) and product recommendation (ShopEasy), powered by a Llama 3.1 8B fine-tuned on 20,000 Nigerian reviews. Built with LangGraph, Langfuse, Pinecone, and Cohere.
A family of neural text-to-speech models bringing natural voice synthesis to Nigerian languages (Yoruba, Hausa, Igbo, and Pidgin), based on Orpheus-3B, enabling voice-first AI for underserved language communities.
Voice-first solar system sizing for Nigeria in five languages, built entirely with sub-4B models for the Build Small Hackathon. Affordable, offline-friendly clean-energy access at the edge.
Safety intelligence platform combining YOLO26-nano real-time PPE detection (mAP50 0.691, 2,540 FPS on CPU) with RAG over OSHA regulations and GPT-4o reasoning for advisory, incident analysis, and live alerting. Deployed on AWS EC2.
Two-stage fire/smoke detection for Oil & Gas facilities pairing YOLOv8s (mAP50 83.4%, 0.5 ms inference on A100, trained on 30,337 images) with a Vision Language Model for context-aware severity assessment and emergency-response recommendations.
Computer-vision prototype for real-time quantification of plastic pollution, with an optimized baseline tuned for the speed/accuracy trade-off. Deployed as a live Streamlit app.
Computer-vision board analyzer and move predictor built during a robotics hackathon, with an integrated preprocessing and inference pipeline that reads a physical board state and recommends the next move.
Anomaly-detection system for IoT/industrial sensor streams, surfacing equipment faults and abnormal conditions early to support predictive maintenance and reliability.
Open-source application that tracks CO2 levels to promote environmental sustainability, collecting sensor data, analyzing it, and providing actionable insights to reduce carbon footprint.
Web application that classifies rail tracks as dirty or leaf-covered to support effective preventive maintenance in railways.
Full-stack pipeline turning corporate documents (PDF, DOCX, PPTX, TXT) into on-demand micro-learning: async Node.js proxy → n8n orchestration → Supabase pgvector RAG with Cohere reranking → GPT lessons → TTS narration → quizzes & flashcards.
End-to-end content engine for a marketing agency managing the full lifecycle from ideation to publication. Integrates Telegram, OpenAI, n8n, Google Docs, Airtable, and social APIs (LinkedIn, X/Twitter, Gmail) to auto-generate, format, and publish SEO content.
AI analytics automation that lets users query massive data warehouses in plain English. Integrates RAG, OpenAI GPT, BigQuery, n8n, and QuickChart to convert natural-language prompts into real-time insights without coding.
Transforms raw sales-form inputs into polished, client-ready proposals within minutes. Integrates n8n, OpenAI, Google Drive, Airtable, and Gmail to create, enhance, validate, and deliver proposals automatically.
End-to-end reporting system that consolidates departmental data from multiple Airtable bases into structured, validated, and visualized reports, updating dashboards automatically on a recurring basis with Make and Airtable.
Automated workflow that processes and validates invoices using AI-driven data extraction. Integrates Make, PDF.co AI Invoice Parser, and the OpenAI API to streamline invoice recognition, extraction, and classification.
AI-powered system that discovers qualified leads, researches their companies, verifies contact information, and writes personalized outreach. Integrates n8n, Apify, Airtable, OpenAI, and Bouncer API for end-to-end prospecting.
Analysis of Airbnb data with sentiment analysis and predictive analytics to surface insights from guest reviews.
Time-series analysis and forecasting model for predicting oil production rates using statistical methods and machine learning.
Analysis of medical-appointment data to identify patterns in patient no-shows and provide actionable insights for healthcare management.
Machine-learning regression pipeline that predicts used-car prices from vehicle attributes, with feature engineering, model comparison, and evaluation.
Customer-review web scraping, sentiment analysis, and a predictive model of booking behaviour, completed for the British Airways data-science virtual experience.
Classification models that predict the stability of a smart electrical grid from simulated operating conditions, with feature analysis and tuning.
Forecasting energy/power consumption from historical demand data using supervised learning, built during the Hamoye data-science program.
Exploratory data analysis and explanatory visualization of a large loan dataset, surfacing the drivers of borrower outcomes and loan performance.
Predictive model estimating gas-turbine Equivalent Operating Hours from plant condition data to support maintenance planning and reliability.
End-to-end data wrangling of the WeRateDogs Twitter archive (5,000+ tweets): gathering from multiple sources and formats, assessing quality and tidiness, cleaning, then analysing and visualising with Python and SQL. (Udacity Nanodegree)
Machine-learning pipeline that predicts age using a hybrid modelling approach, combining engineered features and learned representations for improved accuracy.
Classic supervised-learning project predicting passenger survival on the Titanic, with exploratory analysis, feature engineering, and model comparison.
A collection of computer-vision notebooks and experiments spanning image classification, detection, and preprocessing pipelines.
I design and ship reliable, monitored automations with n8n, Make, and Python, integrating CRMs, payments, databases, LLMs, and communication tools to replace manual hand-offs with intelligent, measurable systems.
End-to-end automations across fintech, HR, customer support, treasury, internal control, risk, and compliance, streamlining operations across multiple subsidiaries (BlackCod Group / Bluebulb).
Agentic workflows for ways-of-work, onboarding, project management, content curation, and community management, cutting manual effort and wasted time by ~60% (CatalyzU, TechStars '23).
Natural-language querying over terabyte-scale warehouses, automated BI dashboards, and recurring report and investor-update generation that turns raw data into decisions without manual effort.
Pipelines that automatically generate, format, and publish SEO-optimized content across LinkedIn, X/Twitter, and email, triggered from Telegram with human-in-the-loop review.
Robust integrations and data migrations across tools and databases (e.g. Firestore to Postgres), built on webhooks and REST APIs for dependable, real-time cross-platform sync.
Reusable templates, automation standards, and monitoring dashboards that keep workflows reliable and scalable across a portfolio of companies, with observability baked in.
Domain-specific Nigerian-context LLM (Llama 3.1 8B, QLoRA) for review simulation. Achieved RMSE 1.114 (vs 1.319 for Claude Sonnet 4) and a 48.5% blind Nigerian-rater win-rate, reaching statistical parity with frontier models at 4× smaller size. Submitted to the DSN × BCT LLM Agent Challenge. Read the paper
Persona-aware product recommendation tuned to Nigerian shopping behaviour, achieving NDCG@10 0.588, the best of 5 models evaluated. Built on 24 culturally-grounded personas across Yoruba, Hausa, Igbo, and Pidgin. Read the paper
Released the persona-conditioned Nigerian review training corpus behind NaijaReviewer-8B, blending real Jumia data with stratified synthetic expansion across register tiers. View on Hugging Face
Performance evaluation of hydrogen production from hybrid Solar-PV electrolysis & chemical-looping biomass gasification with ammonia as an energy vector: exergy, exergoeconomic/exergoenvironmental, and pinch analyses with multi-objective optimisation. (University of Lagos)
Conceptual design of a low-cost, 26,000 TPA biodiesel plant using castor seeds via reactive extraction: Best Undergraduate Design Project Nationwide (NSChE). (Team Design Project, University of Lagos)
Oguibe, F., Makapane, O., Oyelakin, A., & Emmanuel, A. (2025). Published by Clean Technology Hub & Student Energy. Read the report
XGBoost
FastAPI
Zapier
NLTK
Have a project in mind or want to collaborate? Feel free to reach out!