~/timi
~/portfolio

$ whoami_

Timi Owolabi

ML Systems Engineer

I like to think of my work as an ML sandwich. The bottom bread is the runtime infrastructure that models run on, things like inference engines and kernels. The middle, whether that's PB&J, chicken, or whatever filling you like, is the model research, architecting, data prep, and training. The top bread is the cloud infrastructure that monitors and serves those models to millions of people, using tools like Google Cloud, AWS, and Kubernetes. In short, I'm interested in building smart ML models and making them go vroom!

// Lagos, Nigeria

Timi Owolabi, ML systems engineer
01

about

I'm an ML Systems Engineer. Professionally, I work on the infrastructure behind production ML systems. I also spend a good portion of my time on embodied AI research. I lead the Embodied AI Research Group at ML Collective, where we work on generalist robot policies. I’m particularly interested in making generalist robot policies more capable and making them run blazing fast on onboard compute.

I graduated with a 4.85/5.0 GPA in Electrical Engineering from Covenant University. While at Covenant, I led the Robotics Research Team, working on reinforcement learning, vision-language-action models, and low-cost robotic systems. I also worked with Dr. Daniel Omeiza of the Oxford Robotics Institute on autonomous driving and graph neural networks.

I care deeply about teaching and mentorship. As president of AEIES, I founded a student mentorship program to help engineering students develop strong first-principles thinking. I taught machine learning to female engineering students of APWEN. I also founded The Classroom Project, an initiative supporting underserved students with classroom infrastructure and STEM learning resources.

02

skills

// Languages

Python · C++ · SQL

// Machine Learning

PyTorch · TensorFlow · JAX · Hugging Face · MLflow · LangChain · FastAPI · LLM Inference · Model Compression · llama.cpp

// Robotics & Control

Reinforcement Learning · Control Theory · Computer Vision · ROS 2 · MuJoCo · CasADi · Simulation

// Cloud & MLOps

AWS · Google Cloud · Amazon SageMaker · Vertex AI · Docker · Kubernetes · Helm · Terraform · GitHub Actions · Argo CD · GitOps · Karpenter · OpenTelemetry · Grafana

03

experience

professional experience

  1. Jan 2026 - Present

    Platform & AI Engineer (Graduate Trainee)

    Quidax Technologies

    Lagos, Nigeria · Remote

    • Optimized AWS EKS autoscaling with Karpenter, increasing CPU utilization from 20% to 50% and memory utilization from 30% to 70%, reducing infrastructure costs by 8% while maintaining reliability.
    • Modernized legacy manual Kargo deployment workflows by migrating Node, Python, Go, and Java services to reusable GitHub Actions and Argo CD pipelines, helping achieve 98% automated deployment/failover coverage across 1,099 jobs.
    • Instrumented the merchant platform with distributed tracing and logging using OpenTelemetry, enabling engineers to root-cause production incidents independently and reducing cross-team escalations.

  2. Jan 2026 - Feb 2026

    Founding Machine Learning Engineer

    Trotta Technologies, Inc

    San Francisco, California · Remote

    • Served as founding engineer and ML team lead, building the company's AI infrastructure on Google Cloud Vertex AI.
    • Shipped a production multi-agent defense system for phishing, deepfake audio, and synthetic video detection.
    • Built a self-improving threat detection system using LangSmith-hosted agents and anomaly detection networks on Google Cloud Run, reducing verification time from about one hour to under 20 seconds.

  3. Jul 2024 - Dec 2024

    Lead Machine Learning Engineer

    Scrella Technologies

    Lagos, Nigeria

    • Led a three-person ML team and built the company's AI infrastructure on AWS SageMaker, covering the full MLOps lifecycle from data collection to production deployment.
    • Designed an AI-powered phone inspection pipeline using fine-tuned YOLOv11, CLAHE, spectral analysis, and Google PaliGemma.
    • Eliminated manual approval and reduced onboarding review time from five minutes to 18 seconds.

research experience

  1. Jun 2026 - Present

    Embodied AI Research Group Lead

    ML Collective

    Lagos, Nigeria

    • Optimize large world-action models for efficient inference on resource-constrained edge devices.
    • Implemented layer skipping on NVIDIA's Cosmos3-Edge-DROID model to develop the compact SmolWAM world-action model.

  2. Mar 2024 - Sep 2024

    Undergraduate Research Assistant

    WaySense · Oxford Robotics Institute

    Oxford, United Kingdom · Remote

    • Analyzed the Lyft Level 5 dataset for multi-agent vehicle state estimation, lane-change detection, and explainable autonomous driving.
    • Adapted the fast subset-scan algorithm to graph neural networks for statistically significant traffic anomaly detection.
04

projects

Arena memory allocation and reset in the C++ inference engine

LLM inference

2026 - Present

LLM Inference Engine ↗

Timi Owolabi

An LLM inference engine being built from scratch in C++20, starting with zero-copy tensor views, stride-based reshape, transpose and slicing, aligned storage, and arena allocation.

Code ↗
Autonomous quadcopter platform

Embedded flight control

2025

Autonomous Quadcopter ↗

Timi Owolabi

An ESP32 flight controller with modular sensor drivers, cascaded attitude/rate PID control, motor mixing, receiver decoding, telemetry, and GPS/barometer/IMU/compass sensor fusion.

Code ↗
TurtleBot mobile robot simulation

Mobile robot localization

2024

TurtleBot ↗

Timi Owolabi

A TurtleBot simulated in Gazebo with PID motion control and Extended Kalman Filter localization, built on the ROS 2 ecosystem.

Code ↗
CamAI person tracking and action recognition

Real-time action recognition

2023

CamAI ↗

Timi Owolabi, Nelson Elijah, Jesse-Paul Osemeke

An AI-powered camera system that tracks people, recognizes actions in real time, and sends emergency alerts for safety, security, and agricultural applications.

Code ↗
Clothing semantic-segmentation example

Semantic segmentation

2023

ClothSegNet ↗

Timi Owolabi

A U-Net model that identifies garment boundaries and segments clothing items from photographs of people.

Code ↗
05

research

publications & manuscripts

Visual preview for SmolWAM: A Latent World-Action Model for Robotic Manipulation on Edge Devices
NeurIPS 2026 Robotics World Modeling Workshop2026Accepted

SmolWAM: A Latent World-Action Model for Robotic Manipulation on Edge Devices

Timi Owolabi, Hikmah Olawore, Frances Adelakun, Ayotomiwa Oyewumi

SmolWAM is a compact latent world-action model for robotic manipulation, achieving 95.5% success on LIBERO while running 11× faster and using 3.2× less GPU memory than its Cosmos Policy teacher, while retaining future-scene prediction.

06

writings

Oct 5, 20269 min read

Tensors Are Just Lenses ↗

Building the tensor foundations of a C++ LLM inference engine: shared storage, strides, zero-copy views, and memory allocation, explained through animated examples.

07

teaching, leadership, & community impact

Research Intern

Dec 2025 - Present

AI-Powered Assistive Glasses

National Centre for Artificial Intelligence and Robotics (NCAIR)

Developed AI-powered assistive glasses for visually impaired users.

  • Deployed Qwen3.5-0.8B on a Raspberry Pi through llama.cpp for on-device scene understanding and real-time environmental descriptions.

Robotics Research Team Lead

Sep 2024 - Aug 2025

Robotics Research Team

Google Developer Groups on Campus

Led a simulation replication of Stanford's ALOHA bimanual manipulation paper.

  • Developed MuJoCo and JAX vision-based manipulation policies and prototyped a low-cost 3D-printed robot arm.
Students and volunteers during The Classroom Project outreach

Co-Founder

Dec 2024 - Present

Engineering Innovation Empowerment Foundation

EIEF

Hands-on engineering education and support for students with limited access to practical STEM learning.

  • Designed and produced engineering kits for teaching core principles.
  • Organized technical training and mentoring programs.
  • Launched The Classroom Project and delivered classroom desks and STEM learning support to children at the Durumi IDP camp in Abuja.

President

Sep 2024 - Aug 2025

Association of Electrical and Information Engineering Students

Covenant University ↗

Led a 20-member executive team serving more than 2,000 students across projects, competitions, and academic support.

  • Coordinated student projects and engineering competitions.
  • Founded a mentorship program pairing junior students with seniors for academic and career guidance.

Machine Learning Tutor

Aug 2023 - Oct 2023

Machine Learning Education

Association of Professional Women Engineers of Nigeria ↗

Taught foundational machine learning to more than 50 female engineering students.

  • Covered supervised learning, model evaluation, and neural networks.
  • Developed hands-on Python exercises using real-world datasets.

Robotics Instructor

Aug 2023 - Oct 2023

Hands-on Robotics Instruction

Project Cage Technologies

Taught robotics fundamentals and mentored students through complete hardware and software projects.

  • Covered electronics, sensors, actuators, embedded programming, and autonomous systems.
  • Guided system design, hardware integration, programming, and troubleshooting.

Team Lead

Sep 2024 - Aug 2025

Robotics and AI Team

Robotics and Artificial Intelligence Innovators Network

Led applied deep-learning projects and organized model-building competitions for students.

  • Built a CNN-LSTM speech emotion recognition system with PyTorch Lightning.
  • Ran a fruit-classification competition comparing ResNet-18, MobileNet V2, and EfficientNet V2.

Member

May 2021 - May 2023

National Championship Sustainability Projects

Enactus

Contributed engineering work to the national champion team's circular-economy and sustainable-agriculture projects.

  • Designed a soil-nutrient sensor circuit for the Cow Blood to Manure initiative.
  • Contributed to projects turning recycled bottles into road tiles and biomass waste into reusable food packs.

Sustainability Intern

Aug 2023 - Oct 2023

Youth Sustainability Initiatives

Schneider Electric ↗

Combined robotics education with operational support for NGO partnerships aligned with a global youth-empowerment initiative.

  • Ran self-driving-car robotics workshops for young students.
  • Partnered with NGOs on STEM tutoring and reviewed program budgets exceeding $60,000 each.
08

news

  1. Our paper “SmolWAM: A Latent World-Action Model for Robotic Manipulation on Edge Devices” was accepted to the NeurIPS 2026 Robotics World Modeling Workshop.

  2. Muta reached the finals of the Africa Deep Tech Challenge.

  3. The State of Robotics in Africa (SORA) report was released. Check it out here ↗

  4. Our paper Fuzzy-PID Controller for Liquid Level Control of Tank Systems ↗ was published in the NIPES-Journal of Science and Technology.

  5. Graduated with First Class Honours.

  6. Gave the talk “From Campus to Career: Navigating the Engineering Journey” to the Redeemer's University Electrical Engineering Students Association.

  7. Our paper Practical Trade-offs in Neural Network Optimization: Brute Force Search and Gradient Descent ↗ was accepted and published in Engineering Research Express.

  8. Our paper Enhancing Radiological Imaging for Better Healthcare Outcomes Through High Performance Hybrid Approach ↗ was published in the ASRIC Journal on Engineering Sciences.

  9. Gave the talk AI: Our Current Reality and Future Trajectory ↗ to the Nigerian Society of Engineers.

  10. Received a finalist award in the Pan-African Robotics Competition.

  11. Gave the talk Deep Learning: Neural Networks ↗ to the Association of Professional Women Engineers of Nigeria.

  12. Our paper Energy Optimization Algorithm for Reducing Energy Consumption in a Smart Home ↗ was published at IEEE ICMEAS.

  13. Received the Founder's Award for Best WASSCE Result at Wellspring College.

  14. Graduated as valedictorian of the Department of Science at Wellspring College.

09

contact

Collaborations I am interested in are: robotics and embodied AI research and machine learning systems projects.

timiiowolabi@gmail.com