Hey 👋, I’m

Anslem Akadu Anuoluwa

<Data Scientist → ML/AI Engineer>

I build ML/AI systems that don’t just run in notebooks they ship, scale, and actually solve problems. From fine-tuning LLMs for hospitals to deploying career assistants on the web, I’ve learned to take messy data, limited resources, and “what if” ideas , and turn them into production-ready solutions.

My work is Iron Man–style engineering: taking raw ideas, data, and code, and forging them into ML/AI systems that actually ship.

Download CV

About Me

I started my career in Data Science, where I built forecasting and financial models that directly shaped hospital decision-making. That gave me a strong foundation in model development, evaluation, and data-driven insights.

Over time, I realized I didn’t want my work to stop at the notebook or dashboard. I wanted to own the full journey from training ML models to deploying production systems that scale on the cloud and deliver lasting impact.

That push led me into ML/AI Engineering, where I now focus on fine-tuning LLMs, building APIs with FastAPI & Flask, deploying on AWS, and optimizing pipelines with Docker & Bash. I thrive on end-to-end ownership: experimenting, debugging, wrapping models in services, and ensuring they run reliably beyond the lab.

My mindset is simple: ML isn’t just about accuracy scores, it’s about production-ready solutions that solve real problems. Whether it’s predicting hospital finances or building open-source AI assistants, I combine data science intuition with engineering discipline to create systems that last.

At my core, I’m a builder. I learn fast, adapt quickly, and love turning limited resources into working solutions taking “a bunch of scraps” and assembling them into something impactful.

Experience

LifeBank — ML/AI Engineer (May 2025 – Present)

  • Fine-tuned Gemma 1B with QLoRA adapters for hospital finance tasks, including domain adaptation and instruction alignment.
  • Set up and managed AWS EC2 infrastructure for model training, running experiments with bash automation and GPU optimization.
  • Refactored the AI module to migrate from Gemma to a larger Qwen model, improving reasoning and scalability.
  • Designed a multi-model architecture (router, reasoning, and expressive models) for personalized hospital financial assistance.

  • Tech: Python PyTorch Transformers QLoRA Flask Bash EC2

LifeBank — Data Scientist (Financial Systems) (Jan 2025 – Apr 2025)

  • Built predictive models for key hospital financial metrics (profitability, liquidity, leverage, efficiency) using XGBoost and Prophet.
  • Introduced a hybrid-weighting logic (80% naive calc + 20% ML) to improve reliability when real data was limited.
  • Implemented revenue & cash flow forecasting with Prophet, incorporating hospital size and state as regressors for tailored predictions.
  • Tested models with hospital scenarios, surfacing early signals of negative cash flow & financial instability to guide proactive decisions.

  • Tech: Python XGBoost Prophet QLoRA scikit-learn

LifeBank — Data Scientist (Healthcare Forecasting) (Sep 2024 – Dec 2024)

  • Developed healthcare-focused predictive models(NerveX) for blood supply, oxygen availability, and medical stock management.
  • Implemented time-series forecasting pipelines using Prophet and regression methods to reduce risks of stock-outs.
  • Enhanced QuipX system with predictive maintenance models, enabling early detection of equipment servicing needs.
  • Partnered with healthcare teams to translate forecasts into actionable strategies for hospitals and clinics.

  • Tech: Python Random Forest Prophet XGBoost scikit-learn

LifeBank — Data Analyst Intern (Product Analytics) (Jul 2024 – Aug 2024)

  • Analyzed adoption and churn datasets for LifeBank’s healthtech products.
  • Delivered insights on user growth strategies for the Nerve app.
  • Explored adoption and churn datasets for LifeBank’s healthtech products, identifying patterns in user engagement.
  • Delivered actionable insights on user growth and retention strategies for the Nerve app.
  • Crafted a good analysis to communicate product performance trends to the management.
  • Strengthened data cleaning and Python workflows for cleaning and analyzing operational data.

  • Tech: Excel Python Google sheet Data Visualization SQL

Projects

ElevatrAI

Open-source career assistant that helps users identify skill gaps, refine resumes, and generate personalized learning paths in tech.

  • Built an end-to-end Flask web app with clean project structure (modularized `app/`, `resources/`, `data/`, `templates/`).
  • Implemented a skill-to-career recommender system using cosine similarity and NLP-based embeddings (SentenceTransformers) .
  • Integrated with a custom-hosted Hugging Face LLM to generate personalized learning roadmaps.
  • Designed JSON-driven resources for scalability, covering role data, skill mappings, and user session tracking.
  • Deployed first working version on Render and open-sourced the project on GitHub for community contributions.
Live Demo

GitHub Repo

  • Tech:
    Cosine Similarities Sentence Transformer Flask scikit-learn HTML/CSS JSON Render

    • LLM FastAPI Boilerplate

    Production-ready backend service for deploying GGUF language models with FastAPI and Docker.

    • Designed a type-safe API layer with FastAPI and Pydantic schemas for clean request/response handling.
    • Integrated llama-cpp-python for serving GGUF-format LLMs with configurable inference parameters (temperature, top_p, top_k).
    • Implemented production features including CORS, health checks, and OpenAPI documentation for developer-friendly usage.
    • Containerized with Docker for seamless deployment and reproducibility across environments.
    • Deployed on Hugging Face Spaces with working endpoints for /generate, /chat, and /health.
    GitHub Repo

    Live Demo

  • Tech:
    FastAPI Pydantic llama-cpp-python Docker Hugging Face Spaces Python

    • Skills

    Programming & Data:

    Python SQL Pandas NumPy Data Visualization Excel

    Machine Learning:

    Supervised Learning (Classification, Regression) Scikit-learn Time Series Forecasting Feature Engineering Model Evaluation & Metrics Unsupervised Learning (Clustering, Dimensionality Reduction)

    Deep Learning & AI:

    PyTorch TensorFlow NLP LLMs (Hugging Face, QLoRA)

    MLOps & Deployment:

    FastAPI Flask Docker AWS (EC2, S3) Linux & Bash Git & GitHub

    Contact

    Let’s Connect!

    If you have any questions, opportunities, or just want to say hi, feel free to reach out. I’m always open to connecting with fellow professionals, collaborators, and anyone interested in ML/AI.