Engineering · London or the Gulf - Travel to customer sites · Full-time

Applied Optimization Engineer

About 1001

1001 builds AI-powered operational intelligence for the world's most complex, data-heavy environments. We turn fragmented data into a live, unified model of operations and use it to drive better decisions and solve high-stakes problems. Our work sits inside government and large enterprises, in environments defined by critical operations and messy, real-world data.

Our engagements start with forward-deployed teams embedded in the customer environment. They work on real data, build quickly, and iterate until the system proves itself, then scale it across the organization.

The company is backed by Lux Capital, General Catalyst, civ, Hanabi, Sanabil, and 9Yards, with angels including Chris Re, Amjad Masad, Karim Atiyeh, Kareem Amin, and Russell Kaplan.

Working at 1001

We take on high-stakes problems in environments where mistakes carry real consequences. That demands an uncompromising bar, real speed, and systems that hold up under live operations. The people who thrive set that bar for themselves and keep raising it. They own outcomes end to end, bring rigor to everything, and lift everyone around them.

About the role

The optimizer is the core of how 1001 turns operational data into decisions customers can act on. As an Applied Optimization Engineer, you own the math and the code behind it. You take an ambiguous customer problem, often described in the language of logistics, scheduling, or utilization rather than constraints and objectives, and turn it into a formulation that solves and ships.

This is a hacker-scientist role. You need to be strong enough mathematically to reason about formulations, relaxations, and where a model will break, and enough of an engineer to build systems that run in production against real data. The hard part is rarely the textbook version of the problem. It is the version with incomplete data, shifting requirements, and operators who need an answer they can trust.

You will also decide how to solve each problem, not just how to model it. Some problems want a classical solver. Some are better handled by a large language model or a more involved workflow. Knowing which is which, and being willing to throw out an approach that is not earning its place, is most of the job. You will have focused, dedicated time on the optimizer rather than being pulled across product surfaces, and you will help scale our reinforcement-learning-guided optimization work as it moves from promising to production.

What you'll work on

Translate ambiguous customer problems into tractable optimization formulations.
Decide, problem by problem, whether to reach for a classical solver, a large language model, or a more involved workflow, and own the tradeoff.
Scale our reinforcement-learning-guided optimization work from prototype toward reliable production systems.
Calibrate and validate optimization models against real production data, so the output holds up when operators depend on it.
Build with an industrial-strength solver stack and the surrounding engineering, including data handling, integration, and the glue that makes a model usable.
Develop reusable practice, so each new optimization problem starts from a stronger base than the last.

Requirements

3 to 5 years applying optimization to real industrial problems in production, not academic work alone.
Hands-on experience with at least one industrial-strength solver stack such as Gurobi, Cplex, Or-Tools, Pyomo, or Jump, plus strong Python.
Very strong mathematically, with the ability to reason about formulations and know where they will hold and where they will break.
A solid engineer who has shipped production systems, not only built prototypes.
A builder comfortable iterating and pushing through ambiguity to a working result.
A hacker mindset, with strong instincts for where to try, what is likely to work, and when to abandon an approach.

Nice to have