Abdulla Ehsan
Abdulla Ehsan is the founder of Pledge Automations: a data and AI engineer in Dubai who builds AI agents, automations and the studio's own mobile apps.
- Role
- Founder, Pledge Automations
- Based in
- Dubai, United Arab Emirates
- Studied at
- Monash University and The British School in Colombo
Why the studio exists
Abdulla is a data and AI engineer. His work runs end to end: data pipelines and analytics dashboards, AI agents that act inside a company's own tools, and consumer apps on the App Store and Google Play. Pledge Automations is the studio he founded to do that work for other teams.
Every product he builds starts with a problem someone lives with: a team copying data between five tools at 6pm, an enquiry that waits until morning, a gym plan abandoned in week three. He maps why it happens, then builds the smallest thing that fixes it.
The studio is founder-led. The person on your first call is the person who scopes and builds the work, every project has a fixed price agreed before it starts, and clients own all code and accounts from day one.
How he works
The same habits on every project, whether it is an inbox, a dashboard or an app.
Find the real cost
Before any code, measure what the problem costs: hours a week, missed leads, errors. That number decides whether it is worth automating at all.
Build the smallest thing that works
One workflow, end to end, on real data, instead of a six-month plan. You see it working before you are asked to believe in it.
Try to break it
Test against data the system has never seen and cases it was not designed for. If it only works in a demo, it does not work.
Ship it, then watch it
Go live with monitoring, and a human approval step wherever money or customers are involved. Then keep improving what the numbers say matters.
Apps and platforms
Five apps of the studio's own, all live on Google Play and four of them on the App Store, plus platforms built for clients. Each one was designed, built and launched end to end.
TOGA
Live on the App Store and Google Play. First released 2026.
TOGA is a gym and nutrition app with an AI coach: photograph a meal to log its calories and macros, follow a training plan, and ask a coach that reads your own workouts and food log.
Bandwise
Live on the App Store and Google Play. First released 2026.
Bandwise is IELTS speaking and writing practice: answer out loud or on the clock, and an AI examiner marks the answer in seconds against the public band descriptors.
Forge
Live on the App Store and Google Play. First released 2026.
Forge helps people quit vaping, drinking, smoking, cannabis, gambling or any habit they choose, with a clean-day counter, tools for the moment a craving hits and a private AI coach.
Pledge
Live on the App Store and Google Play. First released 2025.
Pledge is a habit and urge tracker built for the hardest ten minutes of quitting: a daily letter from a companion, a voice check-in that untangles what is driving an urge, and one-tap help when it hits.
Quran
Live on Google Play, iPhone version coming soon. First released 2026.
Quran: Memorize, Learn & Audio is a free Quran app with the Arabic, a translation and a transliteration under every ayah, all 114 surahs told as stories, memorization practice and six offline reciters.
reMedy Medicine
Client web platform, delivered 2025.
reMedy Medicine is a revision platform for UK medical students that we designed and built with the reMedy team: a structured knowledge base and curated revision packs behind one calm reading interface.
Research projects
Applied AI research from Monash University, quantitative finance and data visualisation. The code for each one is public.
Perp Quant Lab
Quantitative research
A research and live-trading stack for crypto perpetual futures, built on one rule: a strategy does not count until it holds up on data it has never seen, after fees. Most ideas failed that test. It also checked whether a language model could predict price direction on unseen data. It could not, so the strategy stays systematic.
Built with: TypeScript, statistics, backtesting, exchange APIs
Explainable multimodal sentiment analysis
Final-year project, Monash University
A research prototype that reads a short video three ways, the speaker's face, their voice and their words, then combines them into one sentiment score and shows how much each signal moved it. Nominated for Best Final-Year Project at Monash University.
Built with: Python, RoBERTa, DeepFace, audio emotion model
Sentiment-aware CRM
Group project
A CRM that reads sentiment across email, Slack messages, chat logs and surveys, tracks leads and outreach, and raises an alert when the sentiment of a conversation shifts.
Built with: TypeScript, natural language processing, alerts, analytics
Movie ratings visualisation
Interactive data visualisation
Linked charts that compare films' IMDb ratings, vote counts and Metascores from 2000 to 2017, with a year filter that highlights each release year across the views.
Built with: Vega-Lite
What he works with
Tools chosen for how reliably they will run a year from now, inside the systems a business already pays for.
- AI and automation
- Claude and other frontier language models, agent workflows, and retrieval over a company's own documents
- Data
- Python, SQL, PostgreSQL, MariaDB, Metabase and Tableau dashboards
- Product
- React, React Native, TypeScript, Node.js, Swift, Supabase and AWS
Questions about Abdulla
Abdulla Ehsan is the founder of Pledge Automations, an AI automation studio. He is a data and AI engineer based in Dubai who builds AI agents and automations for businesses, and the studio's own mobile apps.
Yes. Pledge Automations is founder-led, so the person on your first call is the person who scopes and builds the work.
TOGA, Bandwise, Forge, Pledge and Quran: five apps of the studio's own, all live on Google Play and four of them on the App Store. The studio has also built client platforms such as reMedy Medicine, and his research projects cover quantitative finance and explainable AI.
At Monash University, where his final-year project on explainable multimodal sentiment analysis was nominated for Best Final-Year Project.