AI Engineer & NLP Researcher
I build AI that speaks African languages.
AI advocate bridging the gap for more inclusivity for Africans in the AI space. My first-author research on Nigerian Pidgin machine translation won the Didactic Award at ICLR 2025 and I ship the datasets, models, and production systems that put that mission into practice.
My degree is in Biotechnology stretched across nearly seven years by Nigeria's university strikes. While campuses were closed, I taught myself data science, became a Gold Microsoft Student Ambassador (the first from FUTA), and earned six Microsoft certifications, including Azure AI Engineer Associate.
At Data Science Nigeria I found the problem that organizes my work: the languages my community thinks in are nearly invisible to AI. So I built the missing data by hand 2,300 parallel sentences translating biology research into Nigerian Pidgin and published a systematic failure analysis of five translation systems and GPT-4, awarded at ICLR 2025 in Singapore less than two years after I graduated. Today I work at the intersection of research, production AI engineering, and community because inclusion needs all three.
Machine translation & LLM evaluation for low-resource African languages open datasets, honest benchmarks.
Production GenAI: RAG, LLM agents, document intelligence, and AI governance for finance and consumer brands.
Founder of TAG Ecosystem free AI education and community-built language data for Africa.
Led as Principal Investigator at TAG Ecosystem research into glossary-augmented translation that keeps technical terminology accurate when translating into Nigerian Pidgin. Accepted and showcased at the Women in Machine Learning Symposium at ICML 2026 in Korea.
First-author. A systematic failure analysis of five machine translation systems and GPT-4 on scientific text, using a purpose-built 2,300-sentence parallel corpus. Awarded the workshop's Didactic Prize with a spotlight talk in Singapore.
Twice awarded competitive grants by Women in Machine Learning to showcase my work building a Pidgin translation system for education presented as a poster at NeurIPS 2024 in Vancouver, Canada and again at NeurIPS 2025 in Mexico.
Open data for a language AI ignores: Eng-PidginBioData a 2,300-row English↔Pidgin parallel corpus of biological research text, archived with a Zenodo DOI (10.5281/zenodo.18888857) and Eng_PidginEdu, a 26,000+ row educational corpus community-curated with TAG Ecosystem.
AI-powered Know-Your-Customer and document intelligence adopted by Nigerian financial institutions automated identity document extraction and verification that speeds up customer onboarding.
A conversational AI financial assistant that improves financial literacy and makes banking transactions more accessible bringing agentic AI into everyday customer banking.
Anti-Money-Laundering AI for financial compliance intelligent screening and transaction monitoring that helps institutions flag suspicious activity faster and stay regulator-ready.
Brand-activation GenAI image tool selfie-to-branded-art generation deployed as a live microsite experience.
A practical guide to fine-tuning AfriTeVa on Yoruba data the notebook companion to my published afriteva-en-yor model.
RAG with Qdrant vector search, machine translation pipelines, LLM agents, NLP challenges and analytics the full engineering workbench.
Beyond building products, I help organizations integrate AI into their daily workflows and processes: designing agentic AI flows that connect systems end-to-end, deploying copilots (including Microsoft Copilot) into everyday operations, automating time-consuming manual processes, and advising on FinOps to keep cloud AI spend efficient. I've done this inside an FMCG group, for financial institutions, and for teams adopting AI for the first time.
This portfolio isn't just about AI engineering it ships some. My professional profile runs as a live
Model Context Protocol server: add it to Claude, Cursor, or any MCP
client and your AI assistant can query my experience, publications, projects and certifications directly
get_publications, get_experience, contact_flora and five more tools.
Copy the server URL below.
In Claude: Settings → Connectors → Add custom connector → paste it.
Ask your assistant anything: "Has Flora worked with RAG?" it answers from my live profile.
AI agents can also read this site directly via /llms.txt and /resume.json structured, machine-readable, always current.
Vendor-facing AI solutions: I help partners and clients across finance, oil & gas, FMCG and other industries find the right AI solutions and improve their workflows building demos and POCs, then delivering production solutions on cloud environments as an OEM partner with Microsoft, Huawei and Google. Solutions I've built include KYC document intelligence, a financial analyzer, AML screening, a conversational AI safety solution for oil & gas, and AIOps through Microsoft Copilot integration across organizations plus cross-functional FinOps work with the cloud & productivity team, reducing clients' cloud spend.
Built AI-powered solutions helping the marketing team drive product sales the BBNaija 2025 AI Avatar campaign (microsite + Snapchat AR lenses), the Lush Hair AI try-on, AI sales agents automating reporting and campaign insights, and the AI governance framework aligned to the NDPA and EU AI Act protecting consumers' data even while utilizing AI in marketing.
Built Migo's WhatsApp AI bot an LLM-powered conversational system, hosted on AWS, that engages customers differently by segment: past customers still owing, past customers who settled, ongoing customers, and new customers. Beyond customer support, the bot was an insight engine: conversations were stored in a dedicated chat database and I analyzed them to help Migo understand and retain clients surfacing whether defaults were driven by high interest, and whether customers would repay if loans were broken into smaller installments. Delivered through WhatsApp Flows integration, working cross-functionally with the DevOps team on deployment.
AI research showcased at NeurIPS, ICLR and the Deep Learning Indaba. Built the RAG side of GovLLM an LLM enabling government agencies to process and retrieve documents; contributed to AfroLLM, an LLM for African languages; built agentic systems replacing the manual processes of validating the data used to train these LLMs; and led the Nigeria 3MTT DeepTech Ready project aligning trainers and building an LLM grader to fairly grade the program's thousands of learners.
Designed a cross-department data-collection pipeline on Microsoft SQL Server; curated the school's data-science curriculum for its LMS; hosted an SDG-focused hackathon that grew enrolment; analyzed credit-risk data for a student loan initiative.
Managed a 1,000+ student community; wrote a top-viewed Microsoft blog post on data scraping (17,700+ views); built a health chatbot leveraging OpenAI technology at the Game of Learners Hackathon; convened Nigeria's largest Global Power Platform Event with 750 students in attendance.
Microsoft-certified AI Engineer and Data Scientist six Microsoft certifications earned across the Azure AI and data stack, each verifiable below.
Udacity Data Analytics Nanodegree ↗ · Turing College · Dataleum Data Analytics ↗
I deliver AI trainings personally and at organizational level from boardroom upskilling to beginner classrooms. Teaching is where my work started, and it has never stopped:
Organization-wide training at Tolaram Group on practical AI tools (ChatGPT, Gemini, Veo 3, agent workflows) and classes on Data Science Nigeria's AI Everyday program practical AI literacy for professionals across departments.
My teaching started as a Microsoft Learn Student Ambassador running workshops that lifted students' data-analysis understanding by 75% and grew through tutoring roles at UnivelCity Programming School and Axia Africa, turning beginners into job-ready data talent.
As "Flora The AI Girl" I create content demystifying AI and data science for everyday people tutorials, explainers and career guidance across TikTok ↗, Instagram ↗, X ↗, LinkedIn ↗, Medium ↗ and Hashnode ↗ all in one place on my Linktree ↗.
I climbed into AI without a roadmap, a GPU, or an open campus. TAG Ecosystem is that ladder, rebuilt: a free, community-driven ecosystem training Africans into job-ready AI talent and building the open Pidgin data and translation tools that let students learn in the language they think in.
Research led as Principal Investigator at TAG Ecosystem accepted and showcased at the Women in Machine Learning Symposium at ICML 2026.
First-author workshop paper recognized for pedagogical clarity, with a spotlight talk.
Twice awarded competitive grants to showcase my Pidgin education translation research as posters at NeurIPS 2024 (Vancouver, Canada) and NeurIPS 2025 (Mexico).
Applications & Selections Committee for Africa's largest ML gathering at Senegal and Rwanda; Top Research Poster, Indaba 2024.
Selected participant, Sub-Saharan Africa edition (AfriLabs · Meta · Gates Foundation).
Including Women Techmakers events, Pycon Nigeria, DataFest Africa and several others.
