ML & AI

From getting more conversion on John Lewis's Instagram to helping doctors find Pneumonia cases in children's X-ray scans.


Home
ML & AI
Mobile Apps
On-boarding CX
Enterprise SaaS
Website Optimisation
"Transparent Products"
Hackathons
About Me


Past role

Challenge
With Instagram blocking retailers to hyperlink posts to let shoppers get easily to the item's page like they do on Pinterest/Facebook/Twitter, I crafted a cloud-platform that create seamless shopping experience to retailers IG fans - we were the 1st startup in the world to do that.

Adding Machine Learning
After 4 weeks of piloting with the John Lewis Instagram team, we added ML codes to find patterns.

Outcome
After the 2nd month of our pilot, ML results have shown that "cold text" on the post description creates higher engagement rate than "hot text": So no more "This amazing jacket" were the guidelines for JL IG team.

Course Project

Challenge
On the course "AI for PM", I was giving medical data set of 118 X-ray scans. 6 of the scan were labeled as "Healthy" (not having any signals) or have one or two of the Pneumonia signs.

Adding AI
Using Appen, one of the industry’s most advanced AI-assisted data annotation platform, I upload the data, labeled it, crafted the questions the AI needs to "answer" and dealt with fine tuning the project to get for the best results.

Outcome
Got the best grade possible, for this task from the Udacity AI team that reviewed my work.