Machine Learning Data Science

Paul Okafor

+1-405-***-**** # *********@*****.***

https://obinopaul.com ï https://linkedin.com/in/obinopaul § https://github.com/obinopaul Summary

Data Scientist with over 4 years of experience in machine learning, AI engineering, and business intelligence. Skilled in multi-agent systems, fine-tuning LLMs, and developing AI applications. Proficient in Python, PyTorch, Langchain, Power BI.

Education

• Ph.D. in Data Science and Analytics, University of Oklahoma, USA Jan. 2025 – Present

• M.S. in Data Science and Analytics, University of Oklahoma, USA Aug 2023 – Dec. 2024

• M.S. in Petroleum Engineering and Project Development, University of Port Harcourt, Nigeria Nov 2018 – Nov 2019

• B.S. in Petroleum and Gas Engineering, University of Lagos, Nigeria Oct 2012 – Dec. 2016 Technical Skills and Expertise

• Data Science and Analytics: Data analysis; A/B testing; visualization (Matplotlib, Seaborn, Power BI, Tableau); statistical and predictive modeling; OpenCV, YOLO, Detectron2 for image analysis and object detection).

• Machine Learning and AI: Deep learning (TensorFlow, PyTorch), Scikit-Learn, supervised and unsupervised models.

• Programming and Data Tools: Python (Pandas, NumPy), C++, R, SQL for data manipulation and analysis.

• Data Management and Cloud: Data architecture, governance, big data (Databricks, Apache Spark), cloud platforms

(AWS, Azure, GCP).

• Expertise in fine-tuning LLMs (HuggingFace, Ollama, Anthropic, OpenAI), using Langchain and LangGraph for context-aware AI agents, with specialization in RAG systems and multi-agent workflows. Featured Projects

• Pocket Traveler — AI-powered personal travel assistant. Click here to try it (Demo) GitHub Repository Professional Experience

SEQTEK, USA April 2025 – Present

AI/LLM Engineer

• Deployed a production-grade, low-latency multi-agent LLM platform that now serves as the primary interface for all telecom-data queries—built with LangChain + LangGraph orchestration, Azure AI Foundry models (GPT-4.1 & open-source Llama variants), and a Postgres-backed semantic memory layer—cutting manual data-lookup time to near-zero and driving measurable gains in customer-service productivity.

• Delivered a scalable, end-to-end, full-stack conversational AI solution (Node.js UI, containerized backend, Azure) achieving < 5s P95 latency and < 0.5% error rate in live operations.

• Instituted a rigorous evaluation and cost-governance pipeline (latency, factual precision, token economics, regression QA) that guides continuous model iteration; insights from this framework reduced average response cost per query by 38% and validated 97% answer correctness before release. University of Oklahoma, USA August 2023 – Present

Research Assistant

• Conducting AI-driven research on the effects of chemotherapy and cranial radiotherapy on aging in pediatric medulloblastoma survivors, using predictive modeling and clinical data analysis to investigate microvascular damage, cognitive impairment, and potential senolytic therapies to mitigate cognitive decline.

• Developed machine learning models to enhance anomaly detection in manufacturing quality control, reducing misclassification costs and improving process efficiency in collaboration with OU Biotech Core Facility. See paper.

• Collaborated with Cytovance Biologics, USA, to apply interpretable machine learning models for optimizing recombinant protein titer production in E. coli fermentations, resulting in improved yield and reduced protein wastage. See paper. Backyard Innovations Limited, Nigeria September 2020 – July 2023 Data Scientist

• Designed real-time analytics dashboards with Power BI to track energy usage and forecast demand.

• Developed predictive models for smart home energy systems, analyzing client energy usage patterns and forecasting future consumption across different timeframes and areas of the home/offices.

• Identified high-energy usage zones and delivered tailored recommendations to the operations team for efficient smart home installations, optimizing energy efficiency and cost savings.

• Enabled a 20% reduction in operational costs while enhancing client satisfaction by improving both energy efficiency and the aesthetic appeal of smart home setups.

Fortizo Energy Resources Limited, Nigeria February 2020 – July 2020 Technical Sales Engineer

• Created technical bids and proposals, leading to an increase in contract win rate and improved client satisfaction.

• Leveraged technical expertise in process designs, equipment lists, and heat/material balances to enhance deliverable accuracy and efficiency, while cultivating strong client relationships to drive sales targets and revenue growth.

Contact this candidate