Emad Gohari / Toronto, Canada
Practical AI.
Built with care.
Senior Machine Learning Engineer
I work across language models, search, and the ML infrastructure behind them. At Thomson Reuters Labs, I help turn research and domain knowledge into useful AI systems.
From production features to the workflows that help teams improve them.
A few things I’ve worked on
From research to everyday use.
Grounded report summarization
Turning long entity reports into useful summaries, with claims linked to their sources.
In production, processing hundreds of reports daily.Reproducible model retraining
An on-demand workflow that turns research code into reusable, tracked experiments.
An estimated ~40% reduction in researcher time per experiment round.Agent configuration & evaluation
A central place to configure prompts, tools, and agent connections in CLEAR Investigate.
Agent definitions could be reconfigured from one place instead of across many files.On my workbench
Small tools.
Practical problems.
A small tool for a real everyday problem: keeping track of focused work. A Python command-line timer with git-aware sessions, a live terminal dashboard, and a companion macOS menu-bar app.
Explore PomoCLI on GitHub# Start a focused work session
$ pomo start "Build something useful"
# See the day’s work
$ pomo report today
# Open the live dashboard
$ pomo dash
How I approach the work
The model is one part of the system.
Start with the domain.
Work with subject matter experts to understand the question, the data, and what a useful result looks like.
Make the tradeoffs visible.
Compare approaches, inspect failures, and use feedback to improve the system beyond a promising first demo.
Make it repeatable.
Put the data, configuration, and experiment workflows in place so the team can keep developing the work.
Get in touch
Have a problem worth
working through?
I’m interested in useful AI applications, thoughtful engineering, and teams that care about getting the details right.