We work with companies that have algorithmic and operational problems they haven't had bandwidth to attack properly. They need someone who can both think through the problem and ship the solution into their codebase.
Algorithms · Optimization · Operations research
A small team solving hard
algorithmic, optimization, and
operations‑research problems.
Production code in your environment. Not a Jupyter notebook, not a research paper.
Engagements are project-based and scoped to the problem. We can take on standalone work or embed alongside your engineering team. After delivery, we can stay on call to maintain what we built.
Our background is in mobility and logistics, but hard algorithmic problems look similar across industries: marketplaces, fintech, scheduling SaaS, and operations-heavy companies have the same shape of pain.
Selected past work
A sample of recent engagements. Each ended with code shipped into a live system.
Mobility platform
Emergency trip handling
A mobility platform's existing system left ~20% of disrupted trips unrecovered, hurting reliability and rider trust. We rebuilt the algorithm from scratch, redefining how disruption events were classified, matched, and resolved, taking coverage to 98% and exceeding the operational target.
- Trip recovery coverage
- 80% → 98%
- Operational goal was 95%
- Exceeded target
Mobility platform
Search performance & relevance
The platform's search was slow and noisy: long latency, results that didn't match user intent. We profiled the bottlenecks, rewrote the slowest queries and indexing logic, and tightened the relevance scoring. The result: queries ran twice as fast and returned dramatically cleaner results.
- Query latency
- 2× faster
- Irrelevant results
- −90%
Routing engine
Walking-time prediction
The product depended on Google Maps API calls to estimate walking times for routing decisions, costing roughly $0.05 per search at scale. We trained a regression model on historical trip data that predicted walking times with sufficient accuracy to replace the API calls entirely, removing both the cost and the external dependency.
- Per-search API cost eliminated
- −$0.05
- External dependencies
- Zero
Selected problems
Five recurring shapes. Examples, not a fixed menu.
Routing & scheduling
Vehicle routing, dispatch optimization, multi-stop sequencing, time-window constraints. The classic operations research problems, applied to fleets, deliveries, and service teams.
Search & ranking
Query understanding, relevance ranking, result diversity, and the optimization tradeoffs that make search feel fast and accurate at scale.
Matching engines
Two-sided marketplace matching, supply-demand allocation, ride-hail dispatch, recommendation systems where the matching logic itself is the bottleneck.
Resource allocation & capacity planning
Where to place stations, depots, or facilities. How to size fleets. How to allocate constrained resources across competing demands.
Custom algorithms
Most real problems don't fit cleanly into a category. If yours is algorithmic at its core, we can design and build a custom solution, optimized for correctness, performance, and the constraints that actually matter to your operation.
How we work
A predictable shape: short paid diagnostic, scoped build, clean handoff. No surprises.
- 01
Diagnostic
~1 week · paid
We spend about a week with your code, your data, and your team, understanding the problem, sizing the opportunity, and writing up what we'd recommend. At the end you have a written scope with options at different price points, and you decide whether to proceed.
- 02
Implementation
A few weeks · scoped
If you proceed, the implementation is scoped to the problem: usually a few weeks of focused work, with weekly progress check-ins and a clear definition of done. You get production code in your environment and the documentation to maintain it.
- 03
Handoff
Ongoing or clean exit
After delivery, we can stay on call to maintain or extend what we built. Or we step away cleanly and your team owns it. Either is fine.
Who we are
Small, senior, specialist.
Algorism.dev is a small specialist team focused on algorithmic, optimization, and operations‑research problems. Every engagement is led by senior engineers with backgrounds at top mobility startups and tech companies, including Swvl and Amazon. We're deliberately small: projects get our full attention.
- Team size
- Small, senior
- Backgrounds
- Swvl, Amazon
- Specialization
- Algorithms · Operations Research · Optimization
Get in touch
Have a hard problem? Let's talk it through.
Send us a note describing what you're working on. We'll reply with a few questions and see if there's a fit.
contact@algorism.dev