Emmanuel Alabi

Research interests

• Embodied AI: How intelligent agents can perceive, reason about, and interact with physical environments.
• Robot Learning: Learning policies and representations for perception, navigation, manipulation, and control.
• Computer Vision: Visual perception and representation for autonomous systems.
• Autonomous Systems: Systems capable of sensing, planning, decision-making, and action.
• Multimodal Intelligence: Combining vision, language, sensor observations, and action.

Education

• Babcock University

  2019–2023

  BSc (Hons) Software Engineering · Second Class Upper Division, CGPA 4.00/5.00

  Relevant coursework: Algorithms and Data Structures, Discrete Mathematics, Introductory Statistics, Introduction to Operations Research, Artificial Intelligence and Applications, Introduction to Big Data Engineering, Computer Organization and Assembly Language, Operating System I, Database System Design, Implementation and Management

Independent research

• Driver Drowsiness Detection: Reproduction & Evaluation Study

  2026

  Independent Computer Vision Study

  • An oracle-classifier decomposition showed the historical temporal threshold, not residual classification error, dominated false alerts once classifier accuracy was sufficient.
  • Subject-disjoint reproduction on the MRL Eye Dataset; CNN trained from scratch against a HOG+SVM baseline, with deterministic splits, automated leakage checks and three-seed ablations.

  https://github.com/alabiemmanuel177/driver-drowsiness-reproduction

Selected Academic Projects

• BUCODEL: Learning Management System for Open & Distance e-Learning

  2022–2023

  Undergraduate Final-Year Project (SENG490, 72/100), Babcock University

  • Co-designed and implemented a web-based learning management system for Babcock University's Centre for Open and Distance e-Learning, supporting student, lecturer, and administrator workflows.
  • Built course and resource management, assignments, authentication, user administration, collaborative course activity, and role-specific interfaces using React, Node.js/Express, and MongoDB.
  • Conducted requirements analysis, system design, iterative implementation under a spiral methodology, and interface/usability, unit, and integration testing.

Research

• Better Calibration Without Demonstrated Safety Gains: Risk-Calibrated Semantic Navigation Under Distribution Shift

  August 2026 · Embodied AI, Autonomous Systems, Computer Vision, Uncertainty Calibration

  A preregistered ROS 2 and Gazebo benchmark testing whether converting calibrated semantic uncertainty into navigation cost makes a Nav2 robot safer under controlled distribution shift. Calibration improved; navigation safety did not.

  www.emmanuelalabi.com/research/risk-calibrated-semantic-navigation · https://github.com/alabiemmanuel177/risk-calibrated-semantic-navigation

• Early Failure Prediction and Recovery for Mobile Robot Navigation

  August 2026–September 2026 · Embodied AI, Autonomous Systems, Failure Prediction, Uncertainty Calibration

  A preregistered study of whether navigation failure can be predicted early enough to act on. Failures are detectable roughly four seconds ahead, but the learned predictor did not beat a transparent baseline at a fixed false-alert budget, and prediction-triggered recovery did not improve mission completion.

  www.emmanuelalabi.com/research/early-failure-prediction-navigation · https://github.com/alabiemmanuel177/failure-prediction

• Language as an Uncertain Sensor for Robot Navigation

  August 2026–September 2026 · Embodied AI, Multimodal Intelligence, Autonomous Systems, Uncertainty Calibration

  A preregistered study treating a natural-language instruction as a noisy sensor rather than a command. It closed under a deliberately narrowed scope: a frozen upstream threshold made the preregistered calibration mathematically unreachable, and clean-view grounding saturated at zero errors, which removed the negative outcomes the planned models needed.

  www.emmanuelalabi.com/research/language-uncertain-sensor-navigation · https://github.com/alabiemmanuel177/lusn-robot-navigation

Research training

• Research Programme 2026–2027

  August 2026–present

  The programme of study and experimentation I began in August 2026: building the mathematical, perceptual, and control foundations needed to run independent research on embodied agents, and identifying the specific question I want to pursue in depth.

Publications

• Emmanuel Alabi Olasubomi. Better Calibration Without Demonstrated Safety Gains: Risk-Calibrated Semantic Navigation Under Distribution Shift. 2026. [manuscript]
• Emmanuel Alabi Olasubomi. Early Failure Prediction and Recovery for Mobile Robot Navigation: Core Evidence. Zenodo 2026. [dataset]
• Emmanuel Alabi Olasubomi. Early Warning Without Demonstrated Recovery Benefit: Failure Prediction and Guarded Recovery for Mobile Robot Navigation. 2026. [manuscript]
• Emmanuel Alabi. Reproducing a Classic Driver Drowsiness Pipeline: Where the Errors Actually Come From. 2026. [manuscript]
• Emmanuel Alabi Olasubomi. Risk-Calibrated Semantic Navigation Under Distribution Shift: confirmatory evaluation dataset. Zenodo 2026. [dataset]

Selected projects

• Driver Drowsiness Detection: Reproduction & Evaluation Study

  A leakage-safe reproduction of a widely replicated eye-state and temporal-alerting pipeline, built to locate where its system-level failures actually originate.

  https://github.com/alabiemmanuel177/driver-drowsiness-reproduction

• Meta-Harness Plus: Experimental Search Over LLM Agent Harnesses

  A framework for searching over LLM harnesses under a fixed compute budget: Pareto search across accuracy, tokens and latency; successive-halving evaluation; and drop-one ablation to attribute gains to individual components.

  https://github.com/alabiemmanuel177/meta-harness-plus

• BUCODEL: Learning Management System

  A web-based learning management system for Babcock University's Centre for Open and Distance e-Learning, supporting student, lecturer, and administrator workflows across courses, assignments, materials, and class interaction.

  https://github.com/alabiemmanuel177/finalyearproject_backend

• Obelo: Multimodal AI Infrastructure

  A production AI platform combining multimodal generation, retrieval-augmented context, asynchronous orchestration, model routing, and evaluation systems.

Engineering experience

• Founder & CEO, Obelo — AI Systems & Multimodal Content Infrastructure

  Oct 2025 — present
  • Designed and developed a production AI platform combining multimodal generation, retrieval-augmented context, asynchronous orchestration, model routing, and evaluation systems.
  • Multi-agent orchestration over a structured brand representation, with specialised planner, copywriter, image-composer, layout and publisher agents, and a brand-consistency grader checking every output before it ships.

  Multimodal pipelines · Multi-agent orchestration · Model routing · Retrieval-augmented generation · Evaluation systems

• Software Engineer, Punch Group — Production Web Systems

  Jan 2025 — Jul 2025
  • Delivered three production products: GraphQL query optimisation and memory-leak resolution on a Next.js/Express application, a Nuxt.js lead-generation tool with HubSpot and Salesforce integrations, and a real-time support service with in-app video.
  • Diagnosing memory heap exhaustion in a running production application is the closest thing in commercial work to debugging an experiment that fails only at scale.

  Next.js · NestJS · GraphQL · Firebase · Nuxt.js

• Software Engineer, SPay Business — Mobile & Financial Systems Engineering

  Nov 2023 — Dec 2024
  • Worked on production mobile-financial systems, including authentication improvements, application performance, interface redesign, CI/CD automation, and over-the-air mobile deployments.
  • Production engineering maturity: release automation, over-the-air delivery, and performance work under real usage rather than in a demo environment.

  React Native · CI/CD · OTA deployment · Authentication

• Founder, Acadu — Educational Platform

  Feb 2022 — Sep 2025
  • Led product, architecture and engineering: React Native mobile, Next.js web, NestJS backend, PostgreSQL, deployed on Fly.io and GCP with Redis and RabbitMQ for caching, message queuing and real-time event processing.
  • The closest commercial analogue to research infrastructure on this list: queue-backed asynchronous processing, cache invalidation and event delivery across services, with CI/CD on GitHub Workflows. Paused; its architecture informed Obelo's orchestration design.

  NestJS · PostgreSQL · Redis · RabbitMQ · GCP · React Native

• Head of Technology and Digital Strategy (part-time), Royal Gate Group — Technology & Digital Strategy

  May 2018 — present
  • Led technology and digital strategy for eight years: group-wide digital infrastructure and corporate email standardisation, IT and network systems maintenance, and ongoing digital transformation programmes including e-commerce platforms and internal systems across subsidiaries.
  • Sustained ownership rather than project work — eight continuous years, predating and running alongside the degree.

  Systems administration · Networking · Web platforms

• Software Engineer, Union Bank of Nigeria — Enterprise Systems & Security

  Jul 2023 — Nov 2023
  • Built and maintained enterprise applications in React and Java, designed and enforced website security protocols, and delivered merchant notification systems over REST APIs.
  • Working where correctness and security are audited rather than assumed — the closest commercial analogue to a protocol that has to survive review.

  React · Java · REST APIs

Technical skills

Research computing

Python, PyTorch, NumPy, ROS 2, Gazebo, Nav2, Statistical evaluation, Uncertainty calibration, Experiment design

AI systems

Computer vision, Multimodal systems, Retrieval-augmented generation, LLM and model integration, Evaluation systems

Software and infrastructure

TypeScript, SQL, Docker, CI/CD, PostgreSQL, MongoDB, Node.js, React, Next.js

References

• Prof. Stephen Maitanmi

  Professor and Head of Department, Software Engineering, Babcock University

  Supervised the assessed final-year research project (SENG490, 2023).

• Mr. Akinwale Ojo

  Founder and Chief Executive Officer, SPay Business Solutions Ltd.

  Software engineering role, November 2023 to December 2024.

Contact details available on request.