Skip to content
NewSQL, Python and Power BI training for your team
Log in

From Pharmacy to AI Research: The Wuraola Oyewusi Story

How a Nigerian clinical pharmacist taught herself SQL and Python, then led AI research on Nigerian languages.

Muhammad Jamilu Abdullahi 6 min read
Navy cover with the title From Pharmacy to AI Research, The Wuraola Oyewusi story, beside amber and slate bar shapes

The job ad that changed everything

Wuraola Oyewusi was a clinical pharmacist from Ogbomosho, Oyo State, and she enjoyed the work. Then one day, scrolling through job descriptions, she stopped at a data analyst role. She could already do almost everything on the list. Everything except SQL.

So she went to find out what SQL was. That one moment of curiosity took her from the pharmacy into machine learning, then to leading Research and Innovation at Data Science Nigeria, publishing research on Nigerian languages, teaching five LinkedIn Learning courses, and working as a researcher at Imperial College London. Along the way, she started teaching data science in Yoruba.

Where it started: the lab and the pharmacy

Wuraola trained as a pharmacist and began her career in clinical pharmacy. Her training covered microbiology, pharmaceutical chemistry, forensic pharmacy, pharmacology and pharmacy technology, and much of it happened in the laboratory.

She does not describe pharmacy as something she escaped. She enjoyed it, and she says so whenever she tells her story. Her move into data was not a rescue from a bad career; it was a door she noticed and chose to open.

Those lab years still shape how she works. She credits pharmacy labs for her confidence to experiment and her habit of documenting every step. Her clinical vocabulary later let her label health data, find what she needed and ask precise questions.

The turning point: SQL, Python and learning in public

Her first SQL course taught her something simple: data lives somewhere, and there are things you write to get it. From there she found Python, then data science, then AI. She calls the stretch that followed intense, with study across machine learning, programming, mathematics and computing fundamentals.

She did not wait for a new degree. She learned largely from free online material, which she sums up as “the internet is generous.” And she worked where people could see it: she wrote articles about what she was learning, put her projects on GitHub, released open-source health text datasets, and had a contribution merged into pandas, the Python data library analysts use worldwide.

That visible work opened the first door. By her own account, her first data science job offer came after someone read one of her articles. She went on to work with eHealth Africa, a health-data organisation, where her clinical knowledge and her new data skills finally worked side by side.

The breakthrough: leading AI research at Data Science Nigeria

In October 2019, Data Science Nigeria (DSN), the Lagos-based AI community founded by Dr Olubayo Adekanmbi, named Wuraola its Lead for Research and Innovation. A former pharmacist was now in charge of research and publications, DSN's in-house AI products, client consulting and mentoring junior researchers.

Her signature project tackled a problem every Nigerian on social media will recognise. English sentiment tools misread Nigerian Pidgin, where “ginger” means motivation, not a plant, and “tank” means thanks, not a container. Her team built 300 Pidgin sentiment terms that plug into the widely used VADER lexicon, plus 14,000 Pidgin tweets labelled by expert speakers. The enriched lexicon scored Pidgin sentiment better than the original, and the paper was accepted at the AfricaNLP workshop of ICLR 2020, one of the world's leading machine learning conferences.

The work kept coming, most of it rooted in Nigerian life:

WhenMilestone
Oct 2019Joins Data Science Nigeria as Lead for Research and Innovation
Mar 2020Co-author, Improving Yorùbá Diacritic Restoration, released open source with Iroro Orife, David Adelani, Kọ́lá Túbọ̀sún and others
Mar 2020Lead author, Semantic Enrichment of Nigerian Pidgin English for Contextual Sentiment Classification, ICLR 2020 AfricaNLP workshop
Dec 2020Co-author, NeurIPS 2020 workshop paper on crowdsourcing data to track malaria treatment policy at drug outlets in four Nigerian states
Mar 2021Lead author, NaijaNER: named entity recognition for Nigerian English, Pidgin, Igbo, Yoruba and Hausa
Jun 2021Lead author, AFRIFASHION1600, 1,600 labelled African fashion images in 8 classes, CVPR 2021 workshop
Dec 2021Lead author, AFRIFASHION40000, 40,000 GAN-generated African fashion images, Black in AI workshop at NeurIPS 2021
2024Researcher in data science and data curation at Imperial College London; author-instructor of five LinkedIn Learning courses

The malaria paper is where both halves of her career meet. Ordinary residents used low-cost smartphones to mystery-shop drug outlets and check whether they followed Nigeria's first-line malaria treatment policy. It was a pharmacist's question, answered with data.

Impact: Nigerian languages, open data and teaching in Yoruba

Wuraola's research made Nigerian languages and culture visible to machine learning. NaijaNER showed that one combined model could pick out named entities across five Nigerian languages, outperforming separate single-language models, and the models were shared on GitHub. AFRIFASHION1600 filled a gap that global fashion datasets had ignored; a 2025 study still called it the only notable contemporary Afrocentric fashion dataset.

She also widened who gets to learn data science. She started explaining tech concepts in Yoruba on YouTube, and by early 2023 had published about 22 videos watched around the world. Her reasoning: languages that are written, spoken and available online have a better chance of surviving. One viewer told her they never understood statistics until she explained it. Her phrase for it is “curiosity is not language dependent.”

Her teaching now reaches well beyond Nigeria. Her five LinkedIn Learning courses cover hands-on NLP, NLP for clinical datasets, data science and AI for healthcare, Python for health sciences, and data annotation, and she contributed to Springer's AI in Medicine textbook. She says she publishes free content to give back, because free online teachers once did the same for her.

Lessons for aspiring Nigerian data professionals

  1. Read job descriptions as a gap list. Wuraola started by noticing she could already do most of a data analyst role except SQL. Close the gap you can see, then the next one.
  2. Your old career is an asset. Her pharmacy training made her stronger with clinical data and led straight to work like the malaria study. An accountant, nurse or banker brings the same edge to their own domain.
  3. Learn in public. Articles, GitHub projects, open datasets and an open-source contribution built her portfolio, and an article she wrote led to her first data science job offer.
  4. Solve problems close to home. Pidgin sentiment, Yoruba text, Nigerian drug outlets and African fashion took her work to ICLR, NeurIPS and CVPR workshops. Local data is not a limitation; it is often the gap the world has not filled.
  5. Start with free resources. She learned largely from free online material. Paid courses help, but they are not the entry ticket.
  6. Teach what you learn. Teaching in Yoruba and on LinkedIn Learning grew her reach. She applied to LinkedIn Learning through its open instructor recruitment link.
  7. Go deep and keep going. Her own advice to her younger self: keep at it, and go as deep as possible.

From a pharmacy in Nigeria to research at Imperial College London, it started with one question: what is SQL?

Sources

Keep reading

All posts