TOTemitayo Ogunsusi

Seeking Summer 2027 internships

Texas, USA · Ph.D. researcher

Software Engineer / AI Researcher

Building AI systems
that earn trust.

I work where rigorous AI research meets production engineering, turning uncertain models into measurable systems and ambitious ideas into software people can use.

Explore selected work Start a conversation
10,000+AI training & evaluation examples
15%Evaluation precision improvement
2Degrees in electrical engineering
2027Summer internship target

01 / Selected work

Products built for
real life.

End-to-end systems where product thinking, engineering discipline, and applied AI meet.

A connected daily life

Small actions. A clearer picture.

YOUR DAILY LOOPToday.Log. Understand. Grow.
✓
Make room for focusTasks & routines
☺
Check in with yourselfMood & reflection
☾
Find your rhythmSleep & movement
✳
Patterns become perspectivePersonal insights → tomorrow’s routines
↺
Explore the process
  1. Capture tasks, routines, fitness, sleep, mood, and journal entries.
  2. Review activity and history in a connected daily overview.
  3. Explore patterns and AI Coach suggestions, then adjust routines or confirm schedule proposals.
  4. Keep logging to build history and refine the next day. Finance tracking is separate from AI Coach context.

Shipped product · Mobile

01

Routin

An AI-powered life-management app that turns fragmented personal data into a calmer, connected view of daily life.

Read case study
Your next chapter

A direction. Then a way forward.

↗
START WITH YOUExperience + ambition
01 / DISCOVERFind the missing skills
02 / BUILDA path made for youLessons · practice · mentor
03 / ADVANCEYour next role ↗Track progress. Refine your path.
Explore the process
  1. Start with a resume and a target career role.
  2. Analyze experience and compare existing skills with role requirements.
  3. Build a personalized roadmap of stages, milestones, and lessons.
  4. Work through lessons, quizzes, and exercises with AI mentor guidance.
  5. Track progress, reassess skill gaps, and update the roadmap.

AI product · Career intelligence

02

Masar

A full-stack career intelligence platform for turning a resume into an actionable, personalized learning path.

Read case study

Project archive · 08 projects

Earlier work,
still part of the story.

A record of products, platforms, and client work that shaped the way I build today.

02 / Research

Measure uncertainty.
Design for reliability.

Researching how language models reason, collaborate, retrieve, and fail, then building systems that respond intelligently.

Uncertainty-aware collaboration

Know when to ask for more.

INPUTBuild from a task requirement↴
  1. 01Design
  2. 02Code
  3. 03Complete
  4. 04Review
  5. 05Test
u
One question at every stage.Does uncertainty exceed the threshold?
NO →ContinuePass the output forward
YES ↴Retrieve contextInform the next phase
↓ Generated code & testsInspectable from start to finish
Explore the process
  1. A task enters Design → Coding → Code Completion → Code Review → Testing.
  2. Instructor and assistant pairs include CEO / CTO, CTO / Programmer, Reviewer / Programmer, and Tester / Programmer.
  3. At each phase, token log probabilities feed an uncertainty score compared with a phase-specific threshold.
  4. High uncertainty triggers external retrieval. Output and available context pass into subsequent phases.
  5. The demo displays generated code and tests, token streams, retrieval evidence, and heuristic quality checks. It does not execute generated tests.

Preprint · 2026

UA-ChatDev

A multi-agent software framework that detects uncertain intermediate outputs and selectively verifies them before errors propagate.

Agentic AIUncertainty quantificationRetrievalEvaluation
Explore the research
Language meets location

Ask a question. Find its place.

↳ EXAMPLE QUESTION“Which places are nearby?”
SCHEMA + LLM → SPATIAL SQL↗
SELECT name FROM places
WHERE ST_DWithin(…)
◎
PostGIS → matching locationsEvaluate accuracy & query efficiency
Explore the process
  1. Pair a natural-language spatial question with database schema context.
  2. Generate SQL using a language model and schema-aware prompting.
  3. Execute the spatial query against PostGIS to retrieve matching records.
  4. Evaluate execution accuracy and query efficiency against the benchmark.

M.S. thesis · 2025–2026

Geospatial Text-to-SQL

A benchmark and evaluation pipeline for translating natural-language geospatial questions into executable PostGIS queries.

Text-to-SQLPostGISGPT-4Llama 3Evaluation
Explore the research

In the lab

Ongoing research

Ongoing research2026–Present

E²-MAF

Studying how routing, heterogeneous model assignment, and communication pruning can make multi-agent code generation more energy efficient.

Ongoing research2025–Present

AIME-GPT

A campus and conference assistant combining structured content, vector retrieval, and a maintainable administrative workflow.

03 / Experience

Research depth.
Engineering range.

01

Jan 2025–Present · Prairie View, TX

Prairie View A&M University

Graduate Research Assistant

  • Researching reliable multi-agent systems, geospatial Text-to-SQL, retrieval, and energy-aware AI workflows.
  • Built reproducible Python evaluation harnesses for scoring, failure classification, regression analysis, and benchmark reporting.
  • Developed AIME-GPT with PostgreSQL, pgvector, FAISS retrieval, RAG fallback, and content administration.
02

Mar 2024–Jan 2025 · Remote

Turing

Software Engineer / AI Trainer

  • Built Python scoring and red-team test suites; weekly regression reporting improved evaluation precision by approximately 15%.
  • Produced and quality-checked 10,000+ instruction-tuning and evaluation examples.
  • Tested retrieval strategies and standardized prompt chains and typed function-calling schemas.
03

Mar 2022–Mar 2024 · Lagos, Nigeria

SourceMyGadgets

Software Engineer

  • Built a recommendation engine matching customer needs and budgets to suitable products.
  • Delivered marketplace flows across catalog, cart, checkout, order processing, and support.
  • Developed supplier and inventory workflows for a fragmented vendor network.
Temitayo Ogunsusi monogram placeholder

Current headshot
coming soon

04 / About

Engineer by practice.
Researcher by curiosity.

I build reliable AI systems and production software. My work spans multi-agent workflows, retrieval, model evaluation, backend services, and full-stack products, grounded in the belief that useful intelligence must also be measurable, dependable, and thoughtfully shipped.

Education

Ph.D., Electrical Engineering

Prairie View A&M University
Jun 2026–May 2030 (expected)

M.S., Electrical Engineering

Prairie View A&M University
Jan 2025–May 2026

AI systems

LLM agentsRAGTool useUncertaintyRed-teamingEvaluation

Engineering

PythonTypeScriptFastAPINestJSNext.jsReact Native

Data & infra

PostgreSQLpgvectorFAISSRedisDockerAWSCI/CD

05 / Contact

Let’s build something
worth trusting.

I’m exploring Summer 2027 software engineering and applied AI internships, research collaborations, and thoughtful product work.

ogunsusitemitayo99@gmail.com