Skip to content
View jeosol's full-sized avatar
12 followers · 8 following

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report jeosol

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jeosol/README.md

Jérôme Onwunalu, PhD

[Post-Training | ML Systems & Performance | Optimization]

Results-oriented machine learning engineer with several years of experience building optimization and machine learning applications. Proven ability to drive technical vision, mentor engineering teams, and deliver impactful solutions leveraging large language models (LLMs), MLOps, containerization, and orchestration technologies. Experienced in project management, personnel training, development and mentoring of young engineers.

I am focusing on AI/machine learning, optimization frameworks, and high-performance computing. To build better models, I am expanding my knowledge of LLM fine-tuning and alignment techniques through Hugging Face courses: Smol Course , LLM Course and reading related LLM research papers.

🚀 Featured Projects

1. LLM Alignment Evaluation & Optimization Framework (Self study HuggingFace Course: Smol Course)

  • Target Focus: LLM Fine-Tuning, Instruction Tuning, Preference Alignment, DPO, Reinforcement Learning / ML Systems
  • Core Tech: Python, PyTorch, PEFT (QLoRA), Transformers, TRL

2. Cloud-Native Distributed High-dimensional Simulation and Optimization Engine (Large-scale multi-year project)

  • Target Focus: ML Systems & Performance / Stochastic Optimization / High-dimensional simulation
  • Core Tech: Python, Kubernetes (GKE), Docker, Microservices, REST APIs
  • System Capabilities: Some aspects of the project are described in my published paper: "Taming Complexity: Building and Deploying a 270KLOC Integrated Scientific Application" (Presented at ELS'26).
  • Scale: Containerized, multi-threaded worker pools on Kubernetes designed for massive, automated parallel exploration of high-dimensional parameter spaces.
  • Repo. (Note: The project's repository is not hosted publicly due to restrictions, so the link shows a Gource visualization of the repository.)
  • Data analysis Analysis of Docker images build times and Kubernetes (K8s) manifest synchronization times.

📈 Technical Competency Matrix

  • Infrastructure & MLOps: Kubernetes (GKE), Docker, Ray, High-Performance Computing (HPC), Microservices
  • Languages: Python, Go, Common Lisp (Symbolic Knowledge Systems)

📝 Recent Research & Publications (google scholar)

Popular repositories Loading

  1. cl-dnn cl-dnn Public

    Implementation of Deep Neural Networks in Common Lisp

    Common Lisp 1 1

  2. simapi-docs simapi-docs Public

    SIMAPI documents

    Jupyter Notebook 1

  3. llm-post-training llm-post-training Public

    LLM Post-Training, RLHF, PPO, DPO, etc

    Jupyter Notebook 1

  4. heroku-cl-example heroku-cl-example Public

    Forked from mtravers/heroku-cl-example

    Example use of Heroku Common Lisp Buildpack

    Common Lisp

  5. christmas-recipes christmas-recipes Public

    Created to follow Anna Debenham's Tutorial

  6. bakery-store bakery-store Public

    Forked from CloudCannon/bakery-store-jekyll-template

    HTML