Python Engineer, AI and Document Processing
Talon · Amman, Jordan (Hybrid) · Contract, part-time to start, with an equity path
About Talon
Talon is an early-stage startup in Amman, Jordan, building an AI-powered lifecycle management platform for regulatory affairs teams in the pharmaceutical industry. We're starting in Jordan and expanding across the wider MENA region, including Egypt and Saudi Arabia.
Regulatory affairs is still a document-heavy discipline, and much of the work is done by hand. Automating it reliably is a challenging problem in applied AI and document engineering, one that demands careful handling of document structure, layout, evidence, and accuracy.
Our first product is in active development, with a working system already in place. Now we're looking for an experienced engineer to assess it and help shape what comes next.
The first engagement
A short, paid technical assessment. Approximately 15 to 20 hours over two weeks. You'll start by getting familiar with the existing codebase and assessing its technical foundations.
The initial scope:
Install and run the application on your own machine using only our setup documentation, noting anything that breaks or isn't explained. Then review the Python codebase, identifying what's working well, what's fragile, and what needs improvement. Assess the architecture, code quality, testing, dependencies, documentation, and overall readiness for deployment. Deliver a practical technical assessment outlining your findings, priorities, and recommendations.
This is a fixed-fee project, agreed upon before starting, with payments made weekly throughout the engagement.
What it could lead to
If we work well together, there's an opportunity to take a much larger role in building Talon's engineering foundations and product.
The immediate priorities are preparing the platform for enterprise and cloud deployment, extending it across more of the regulatory lifecycle, and building the technical foundations for new regulatory workflows and new markets across the region, all while improving reliability, performance, and document processing. For the right person, we're also open to discussing an equity with vesting.
Who we're looking for
We're looking for someone intellectually curious, creative, disciplined, and open-minded. Someone who enjoys thinking through problems, exploring possibilities, and engaging with the reasoning behind how things are built.
You should be comfortable moving between the conceptual and the technical. We value someone who can discuss the underlying logic of a system, challenge assumptions, consider alternative approaches, and translate those ideas into sound engineering decisions.
You should also be someone who takes genuine ownership, moving from an idea or problem to a working solution. You approach unfamiliar problems creatively, contribute your own ideas, and are comfortable both disagreeing and changing your mind when presented with a better argument.
AI-assisted engineering
We use Claude Code in our development workflow and expect strong practical proficiency with it. However, Talon follows a strictly human-led engineering process. Architecture, reasoning, implementation, and review remain the engineer's responsibility at every stage.
What we're looking for
3+ years of hands-on Python development in production environments.
Strong experience with PDF processing, including text and layout extraction, tables, and OCR.
Practical experience building applications with LLM APIs.
Demonstrable proficiency with Claude Code and AI-assisted development workflows.
Strong software engineering fundamentals, including architecture, testing, code review, and version control.
Ability to explain technical decisions and problems clearly to non-technical colleagues.
Based in Amman, or able to work on-site with customers here when needed.
Less experience, but work that shows otherwise? Apply anyway.
Interested?
Send me a message with your GitHub profile or a code sample, along with a few lines about the most challenging document-processing problem you've worked on.
We'd also be interested in hearing how you use Claude Code in your development workflow, how you approach code review, and how you maintain technical documentation.