- Led the development of the Analytics Engine, a pivotal component driving diverse data analyses, from descriptive statistics to advanced machine learning for LLMs.- Engineered the Data Language Model (DLM), a natural language query interpreter, transforming user queries into actionable commands, and enhancing the platform's usability.- Designed and implemented Retrieval-Augmented Generation (RAG) systems. Leveraged vector databases and similarity search algorithms to optimize retrieval performance and reduce and eliminate hallucinations. - Successfully built and tailored Private Large Language Models, ensuring private, secure, and efficient generative AI capabilities, aligning with specific industry contexts and user preferences.- Applied Parameter Efficient Fine-tuning (PEFT) techniques, including LoRA and QLoRA in fine-tuning large language models for data analytics reporting, reducing computational costs.- Facilitated integration with key data sources, including Adobe Analytics, Google Analytics, and CSV documents, enhancing the platform's versatility in handling diverse data formats.

- Systematized the transformation of clients' business challenges into well-defined machine-learning problem statements, ensuring a formalized approach to problem-solving.- Articulated comprehensive machine learning data prerequisites for clients, guaranteeing the meticulous collection of high-quality data indispensable for the effectiveness of machine learning processes.- Developed credit default predictive models utilizing diverse machine learning tools hosted on the Microsoft Azure platform, specifically engineered to forecast employee attrition for a prominent commercial bank in Africa.- Constructed a robust credit scoring model employing PySpark, adeptly handling extensive big data sets encompassing millions of customer records, to enhance credit risk assessment processes.

- Orchestrated data aggregation from diverse financial sources using the Mono and Okra API.- Executed feature engineering techniques, including Term Frequency-Inverse Document Frequency (TF-IDF) vectorization, to seamlessly convert unstructured textual data into numerical features.- Leveraged KMeans clustering, to effectively group transactions based on feature vectors, providing insights for decision-makers and facilitating robust credit risk assessments.- Conducted result analysis, categorization, and statistical analysis to unveil pattern and to deliver actionable insights for individuals, empowering them to make informed financial decisions with confidence.- Deployed the model using Python FastAPI library

- Built Buy-Now-Pay-Later Product recommender system for Credit application mobile app.- Built an end-to-end customer credit default ML model with explainability and interpretability insight.- Developed intelligence-sharing dashboards, providing company-wide access to collected data for credit risk assessment.- Designed customized data monitoring and reporting dashboards for the management to monitor team performance

- Developed a Visitors Management System with C#