Ibrahim Fadhili

AI Engineer and Data Scientist with 3+ years of experience designing, developing, and deploying AI-powered applications across analytics, machine learning, Generative AI, and intelligent automation. Experienced in building end-to-end AI systems including machine learning models, Retrieval-Augmented Generation (RAG) pipelines, LLM-powered applications, AI agents, and scalable intelligent software systems, with a strong focus on AI security, governance, and responsible AI development. Skilled in Python, PyTorch, Scikit-learn, REST APIs, vector databases, and modern AI frameworks for developing production-ready AI solutions.

Passionate about solving real-world problems through intelligent software, with experience across healthcare, finance, enterprise analytics, and knowledge retrieval systems.

Lead AI Engineer - AgileBusiness Solutions Limited

(2024-04)

• Architected end-to-end data ingestion pipelines capable of processing structured data, documents, images, audio, and video into unified knowledge bases for AI-powered applications.
• Designed knowledge isolation architecture that securely restricts each client to its own private knowledge base while enabling scalable Retrieval-Augmented Generation (RAG).
• Designed and deployed a self-hosted vLLM + Ollama inference infrastructure, exposing internal REST APIs that significantly reduced reliance on proprietary AI providers and lowered inference costs.
• Fine-tuned and optimized open-source language models using proprietary organizational data to improve domain-specific performance and enterprise retrieval accuracy.
• Integrated AI capabilities into multiple enterprise software solutions, enabling intelligent document search, workflow automation, analytics, and conversational assistants.
• Defined and executed the company's AI strategy, establishing technical standards, governance, and implementation roadmaps for AI initiatives.
• Led the organization's AI Research & Development initiatives, evaluating emerging models, frameworks, and deployment strategies for enterprise adoption.
• Championed the adoption and customization of open-source AI technologies, reducing infrastructure costs while improving flexibility, privacy, and vendor independence.
• Collaborated with engineering, product, and business teams to identify AI opportunities and deliver scalable production-ready solutions.

Chief Data Officer - MsingiAI

(2024-04)

• Co-led the research and development of Swahili-first AI models for speech and language technologies, focusing on improving accessibility for African languages.
• Curate, clean, and preprocess large-scale speech and text datasets for model training, validation, and evaluation.
• Developed and optimized Automatic Speech Recognition (ASR) models for accurate speech-to-text transcription across diverse Swahili dialects and real-world audio conditions.
• Design and train machine learning and deep learning models for multilingual Natural Language Processing (NLP) tasks and Fine-tune open-source foundation models for African language understanding and generation.
• Evaluate model performance using objective metrics such as Word Error Rate (WER), Character Error Rate (CER), BLEU, and Mean Opinion Score (MOS) improving transcription accuracy across target Swahili dialects.
• Experiment with transformer-based architecture and modern speech AI frameworks to improve model accuracy, robustness, and inference efficiency.
• Collaborate with researchers and engineers to develop scalable AI solutions supporting education, accessibility, and digital inclusion across Africa.
• Contribute to the development of reusable machine learning pipelines, model evaluation frameworks, and deployment workflows for production AI systems.

AI Data Annotator (Computer Vision & Video QA) - Meta AI - Remote

(2023-10 - 2025-12)

• Evaluated large-scale multimodal AI models (vision + language) for safety, reliability, and alignment in high-traffic systems
• Developed evaluation methodologies and failure-case analysis for generative outputs across vision, audio, and text.
• Identified edge cases and model vulnerabilities, contributing to safer deployment decisions and model hardening
• Contributed to safety and alignment validation for AI Systems serving millions of users.

Open-Source AI Developer - Graph Fusion AI

(2023-11 - 2025-04)

• Developed AI systems leveraging knowledge graphs to improve contextual reasoning, entity relationships, and enterprise knowledge retrieval.
• Designed graph-based retrieval architecture that combined structured knowledge with Large Language Models to enhance response accuracy and explainability.
• Built GraphRAG pipelines integrating graph databases with semantic retrieval for complex multi-hop reasoning tasks.
• Designed experimentation frameworks to evaluate graph traversal strategies, retrieval quality, and LLM grounding.

Master of Science - Financial Engineering - WorldQuant University (2024-10)

BSc - Actuarial Science - Kisii University (2019-09 - 2023-12)

Diploma - Business Administration - Tekedia Institute (2022-01 - 2022-12)