ML/AI Engineer & PI. Building a Nigerian chest X-ray benchmark to measure how well Western AI models work on African patients.
Joined March 2019
- Tweets 53
- Following 216
- Followers 27
- Likes 543
4 Photos and videos
Pinned Tweet
Mar 27
Building Africa's first chest X-ray AI benchmark from Nigerian hospital.
What happens when AI trained on American chest X-rays is deploy in a Nigerian hospital? No one has rigorously measured how well they work on our patients.
I'm 9 weeks into a research project to find out.
1
43
Apr 3
Week 10/48. This week I audited our entire NLP pipeline.
The question: should you use rules or transformers for medical text?
The answer surprised me. Thread.
1
1
12
Apr 3
Also verified negation detection across 5,517 reports.
0 failures out of 134 negated mentions.
And reviewed 617 reports with 3 labels — 89.9% of all label assignments supported by text evidence.
1
1
8
Apr 3
The lesson: don't choose rules OR transformers.
Use both. Rules give you interpretable precision. BERT gives you semantic reach.
Our ensemble: rules primary on 11 labels, BERT primary on 2.
495 reports ready for radiologist validation.
#MedicalAI #NLP #ChestXray #Africa #BuildInPublic
1
12
Software retweeted
Mar 28
A prompt you can use in Claude to help build & launch apps, v/@milesdeutscher.
16
291
1,574
202,720
Mar 27
The pipeline: two extractors working in parallel.
Rule-based: 447-phrase bilingual dictionary with negation detection
BioClinicalBERT: fine-tuned transformer, Normal as 14th label
Rules handle 11 labels. BERT handles 2 that rules can't detect.
They vote. Flag disagreements.
8
Mar 27
The dataset: 5,517 chest X-rays from FMC Ebute-Metta, Lagos.
Free-text radiology reports. English mixed with local medical terms. 48% had no section headers.
Step 1: turn those reports into structured labels 13 pathologies Normal.
That took Weeks 1–8.
1
64
Software retweeted
26 Mar 2023
Excitement is brewing as I gear up to witness #StudentSummit event. Can't wait to connect with like-minded individuals and dive deep into the world of innovation and creativity!!!
Thanks to @MSFTReactor @MicrosoftNG @MicrosoftLearn @TheOyinbooke
@japhletnwamu #SSLagosWatchParty
12
7
25
2,240
Software retweeted
26 Feb 2023
Recite and Retweet
104
7,505
25,606
1,148,191
Software retweeted
31 Jan 2023
Reminder
1,057
14,337
86,225
11,123,335
