As a machine learning researcher, I focus on reducing the high cost and energy footprint of sophisticated GPUs by designing scalable, privacy-preserving architectures. My work emphasizes building resource-efficient AI systems with a focus on trustworthiness, privacy, and real-world applicability. I enjoy exploring human neural capabilities to design sensitive, aware, and efficient algorithmic methods, and I actively seek collaborative opportunities to translate research into practical, ethical AI solutions. I am passionate about creating scalable ML systems, advancing multimodal and privacy-preserving techniques, and mentor-ing others through hands-on research and community initiatives. I value clear communication, rigorous experimentation, and open science, and I strive to make advanced AI accessible and trustworthy for diverse users and applications.

Samuel Oyeneye

As a machine learning researcher, I focus on reducing the high cost and energy footprint of sophisticated GPUs by designing scalable, privacy-preserving architectures. My work emphasizes building resource-efficient AI systems with a focus on trustworthiness, privacy, and real-world applicability. I enjoy exploring human neural capabilities to design sensitive, aware, and efficient algorithmic methods, and I actively seek collaborative opportunities to translate research into practical, ethical AI solutions. I am passionate about creating scalable ML systems, advancing multimodal and privacy-preserving techniques, and mentor-ing others through hands-on research and community initiatives. I value clear communication, rigorous experimentation, and open science, and I strive to make advanced AI accessible and trustworthy for diverse users and applications.

Available to hire

As a machine learning researcher, I focus on reducing the high cost and energy footprint of sophisticated GPUs by designing scalable, privacy-preserving architectures. My work emphasizes building resource-efficient AI systems with a focus on trustworthiness, privacy, and real-world applicability. I enjoy exploring human neural capabilities to design sensitive, aware, and efficient algorithmic methods, and I actively seek collaborative opportunities to translate research into practical, ethical AI solutions.

I am passionate about creating scalable ML systems, advancing multimodal and privacy-preserving techniques, and mentor-ing others through hands-on research and community initiatives. I value clear communication, rigorous experimentation, and open science, and I strive to make advanced AI accessible and trustworthy for diverse users and applications.

See more

Experience Level

Expert
Expert
Expert
Expert
Expert
Expert
Expert
Intermediate

Language

English
Fluent

Work Experience

Machine Learning Engineer (Contract) at SDT Corp
January 1, 2025 - January 1, 2025
Fine-tuned LLaMA-3.1 on robotics data; memory optimization with minimal loss; integrated into a chatbot pipeline with knowledge graph functionality and RAG flows.
Co-Lead Investigator at ML Collective
January 1, 2025 - November 12, 2025
Privacy Isn’t Free: Benchmarking the System Cost of Privacy-Preserving ML. Advisory: Steven Kolawole (CMU). Contribution: Developed PrivacyBench, benchmarking privacy-utility-cost on CNN and Transformer medical image datasets; highlighted energy consumption implications.
Python Engineer Intern at Vale Finance Limited
January 1, 2024 - January 1, 2024
Developed a telecom service provider classification model with 96% accuracy; integrated backend with FastAPI; streamlined training and deployment pipelines.
Contributor at ABC Project (Multimodal Agreement-Based Cascading)
January 1, 2024 - November 12, 2025
Experimented with SG Lang inference library using different multimodal LLMs (1B to 40B parameters) on the MMLU benchmark to evaluate agreement-based cascading efficiency.
Co-Lead Investigator at Cohere Labs / Aya Expedition Project
January 1, 2024 - November 12, 2025
Secure and Scalable Federated Learning for Bank Fraud Detection; designed HFL transformer architectures and applied them to the BA F-base dataset, outperforming several baseline techniques.
Co-Author at Aya Expedition Project
January 1, 2024 - November 12, 2025
Co-authored Sparse Upcycling Aya Vision; built evaluation pipeline for the Multimodal LLaVaBench; aim to reduce training FLOPs and improve convergence speed for multimodal vision-language models.
Contributor at Multimodal Agreement-Based Cascading (ABC)
January 1, 2024 - November 12, 2025
Investigated ABC-style multimodal cascading approaches using SG Lang/LLM reasoning to improve cross-modal alignment; evaluated efficiency on benchmark tasks.
Co-Lead Investigator at Rebus Puzzle LLM Research
January 1, 2024 - November 12, 2025
Understanding LLM Reasoning Capabilities through Rebus Puzzles. Annotated and experimented on visual language models (VLMs) across tasks from zero-shot to few-shot to assess reasoning and applicability for idiomatic rebus puzzles; evaluated both proprietary and open-source models.
Co-Lead Investigator at Bank Fraud Detection Project
January 1, 2024 - November 12, 2025
Secure and Scalable Federated Learning for Bank Fraud Detection. Designed HFL transformer architectures and applied the techniques to the BAF-based dataset; outperformed the majority of techniques on the BAF benchmark.
Independent ML Researcher at ML Collective
January 1, 2024 - November 12, 2025
Co-led and contributed to multiple ML research projects focusing on privacy-preserving machine learning, energy efficiency, and scalable AI architectures. Developed the YAML-based PrivacyBench framework to benchmark privacy-utility-cost across CNN and Transformer medical image datasets, yielding insights into energy consumption and privacy trade-offs.
Co-Lead Investigator at ABC Project (Understanding LLM Reasoning Capabilities through Reb us Puzzles)
January 1, 2023 - November 12, 2025
Annotated and experimented with visual language models on a range of tasks to study reasoning capabilities of proprietary and open-source models.
Core Team Community Lead at Google Developer Student Club, FUNAAB
January 1, 2022 - January 1, 2022
Organized online and offline events, workshops, and bootcamps for over 3,000 students; built community programs to foster ML/AI skills and engagement.

Education

BSc in Computer Science at Federal University of Agriculture, Abeokuta
January 1, 2019 - January 1, 2024
Bachelor of Science (BSc) in Computer Science at Federal University of Agriculture, Abeokuta
January 1, 2019 - January 1, 2024

Qualifications

Best Poster Award
January 1, 2025 - November 12, 2025

Industry Experience

Software & Internet, Media & Entertainment, Education, Financial Services, Professional Services, Other