Opeyemi Bamigbade

AI/ML Engineering

I build the runtime and platform layers that move AI from a model into a dependable system: inference, orchestration, observability, and operational workflows.

Selected systems

VAAS visual anomaly analysis

Research in production context

VAAS inference library

Packaging Vision-Attention Anomaly Scoring as a Python library, with pretrained model loading and a documented pipeline for image integrity analysis.

Computer vision
Python library
Model inference
  • Loads versioned pretrained checkpoints through a reusable inference pipeline.
  • Returns global and patch-level anomaly scores with a visual anomaly map.
  • Connects the published method to repeatable experiments and downstream workflows.
100+ monthly downloads 5+ model versions 26+ GitHub stars
Hugging Face PyPI Source code
Platframe AI workflow platform

Operational AI platform

Platframe

A governed workflow platform for operational investigations, designed around explicit lifecycle stages, bounded AI assistance, evidence handling, and replayable history.

Runtime coordination
Workflow governance
Operational replay
  • Separates case lifecycle, evidence analysis, model execution policy, governance, and analyst assistance into bounded subsystems.
  • Records the decisions and events needed to reconstruct an investigation after execution.
4 bounded runtime subsystems 3 human-in-the-loop components 1 correlated operational replay
Live demo Source code System walkthrough
Inference engine comparison from Inferex

Inference engineering

Inferex

A hands-on LLM inference engineering stack for testing gateway design, quantisation, serving engines, observability, and GPU deployment.

RTX 5090
WSL2
vLLM
SGLang
  • Compares engine and quantisation behaviour under measured workloads, including latency, throughput, and GPU memory.
  • Runs the gateway, inference engines, metrics, and dashboards together in a local Kubernetes environment.
  • Documents where hardware and environment constraints affect the benchmark results.
2 serving engines compared 4 quantisation variants tested 32 GB GPU memory
Source code System walkthrough

Industry experience

Lead AI Engineer · Curacel

2021–2023

Led a five-person team delivering the company's first production AI system for motor-insurance claims. Built computer-vision services for damage localisation and image classification, an OCR-driven claims-processing pipeline, and cloud MLOps workflows across AWS and GCP. Also integrated vector retrieval into medical-claims workflows.

9% improvement in claims throughput

AI & Backend Engineer · Philanthrolab

2021–2022

Designed NLP-based eligibility scoring and recommendation workflows for social services, taking research prototypes into deployable services. Built the supporting APIs, data pipelines, user-behaviour tracking, and integrations that connected models to product workflows.

Machine Learning Engineer · Clinify

2020–2021

Designed and evaluated machine-learning components for healthcare applications with product and engineering teams. Supported system-level testing and integrated deployed models into backend services.

Machine Learning Engineer · Data Science Nigeria

2019–2020

Built applied ML and NLP systems spanning ranking, conversational agents, and computer vision. Deployed resource-efficient models on constrained hardware, including Raspberry Pi, and developed modular services for client integration.

Skills and areas of practice

Programming and frameworks

Build across model development, data processing, and service integration.

Python · Go · PyTorch · TensorFlow · CUDA · Scikit-learn · OpenCV

Modeling and representation learning

Develop and evaluate computer-vision and multimodal methods.

Vision and vision-language models · Representation and contrastive learning · Transformers · Self-supervised and few-shot learning · Generative models

Inference and optimization

Move models into efficient, measurable inference workflows.

vLLM · LiteLLM · ONNX · TensorRT · TVM · Hugging Face · OpenCLIP · DeepSpeed · Quantization

Systems and deployment

Build the APIs, cloud infrastructure, and operational practices around ML services.

FastAPI · Flask · MCP · Docker · Kubernetes · AWS · GCP · CI/CD · MLOps · Monitoring