Available Planning · scheduling · packing software

Better results from your code, proven before you merge.

For software teams whose product has a number for “better”.

Lower cost, faster answers, better plans. Argmax tries ideas to improve your code and keeps only the ones that work. Your best engineers get their time back, and nothing changes until it is proven better and you approve it.

Not a general coding assistant. It needs a number to improve.

What you get from every try

Proven better
Old and new code run on the same problems, several times. A gain only counts if it is real, not luck.
You decide
Every win arrives as a pull request with its results. Nothing merges without your review.
Nothing wasted
Ideas that fail leave notes behind, so the next idea starts smarter.
Yours to keep
All code and all notes live in your own repo. You keep everything, always.
What it costs you today

Your team knows how to make it better. There's never time to prove it.

Nothing breaks. Your product just stops getting better.

01

Each idea takes weeks.

Someone reads up, builds it, runs the tests and compares the numbers. One idea can cost a senior engineer weeks, and the list of ideas keeps growing.

02

Gains quietly disappear.

A change looked good on a few tests. Two releases later the gain is gone, and nobody knows when. “It seemed faster” is not proof.

03

The cleanup never happens.

Everyone knows the core needs a cleanup. Nobody can show it won't slow things down, so it waits. The code that matters most gets the messiest.

04

Know-how leaves with people.

Why is that setting 0.37? Why was that rule removed? The answer was in someone's head. When they leave, you keep the code but lose the reasons.

Each of these is small. Together they turn a product that leads into one that just keeps up.

Q1  two ideas started
Q2  one shipped, one dropped
Q3  the shipped one quietly slips back
Q4  its author changes teams
→  net result for the year: unknown
What changes

What you get instead.

Argmax does the slow work and keeps only what it can show is better.

Before: Weeks of senior time per idea

Your engineers get their time back

AI studies your code and each idea, writes the code and measures it. Your team reviews results instead of producing them.

Before: “It seemed faster”

Gains you can trust

Old and new code run on the same problems, several times each. A change counts only if it is clearly better, not lucky, and nothing else got worse. A fixed check decides, not the AI.

Before: A cleanup nobody dares to start

Cleanups that are safe to merge

A cleanup only goes through if every test passes and nothing gets slower or worse.

Before: Failed ideas vanish; know-how leaves with people

Every failure makes the next try smarter

When an idea fails, its code is thrown away but its notes stay in your repo. The next idea starts from what was learned.

Log of triesexample
012 tune a setting to the number of machines dropped
013 new rule for which job goes next dropped
014 reorder the jobs that set the finish time kept
015 same, and favour jobs that are due soon kept
Why try 013 failed, passed on to try 014: “the rule for picking jobs isn't the problem; look at the jobs that set the finish time.”

The failed try is what pointed the next one the right way.

How it works

Five steps. You can see every one.

Each step leaves notes in your repo and updates your ticket, so you can always check what happened and why.

  1. 01
    You or Argmax

    Add an idea

    An idea is a normal GitHub ticket. Your team and Argmax share one list.

  2. 02
    AI

    Read up

    It studies your code and the idea, then writes a plan. It also looks at why the last idea failed.

  3. 03
    AI

    Build and measure

    It writes the code, then builds, tests and measures it on a separate copy of your code.

  4. 04
    Fixed check

    Decide

    Plain code compares old and new on the same problems. No AI gets a vote.

  5. 05
    You

    Review or learn

    A win becomes a pull request for you to review. A failure leaves notes for the next idea.

An honest judge

The AI suggests and builds. The yes or no comes from a fixed check your team can read and run again.

Your main code stays untouched

Every try runs on a separate copy. Only wins reach you, and only when you merge them.

Seven kinds of improvement · each with its own check
KindWhat it changesCheck
search Smarter ways to look for better answers clearly better results
settings Tuning the settings you already have clearly better results
rules Better rules for building and choosing answers clearly better results
cleanup Tidier structure, same results all tests pass · nothing worse
new rules Handling a new kind of business rule passes new test problems
splitting Breaking big problems into smaller ones passes new test problems
inputs New inputs and outputs passes new test problems
Is it a fit?

You need four things. With all four, Argmax can improve your code.

A number for “better”

Cost, service level, run time, or how close you get to the best answer.

Test problems

A set of examples that looks like the work your customers give it.

A build that runs itself

One command that builds and tests the code, with nobody clicking.

A fair comparison

Old and new code can run on the same examples.

No number to improve? Then Argmax isn't for you. It is not a general coding assistant.

Proof

Don't take our word for it. Look at the history.

Trial access

See every try: the wins and the failures.

Argmax is working on a trial program that plans jobs on factory machines. Ask for access and see each idea, why it won or failed, and the changes that came out of it.

  • Every win is a pull request with its results
  • Every failed idea keeps its notes
  • It all lives in the repo: copy it and you have the full history
Request trial access →
Best result so farillustrative · shape only

Your best result only moves up when a change wins.

result ↑ tries →
Not customer data. It shows the idea: a failed try never makes your result worse.
Built on published research
  • FunSearch: AI-written packing rules beat the classic ones
  • EoH: AI improves its ideas and its code together
  • ReEvo: AI learns from written feedback on past tries
FAQ

Practical questions

What if it only gets good at our test problems?

Honest answer: the result is only as good as your test problems. Argmax makes that visible instead of hiding it. Old and new code run several times on the same problems, so one lucky run doesn't count. You can lock the test setup so no try can change it. New features must come with new test problems. And your set of tests grows as your code does. If your tests don't look like real customer work, that's the first thing we'll talk about.

Who owns the code?

You do. Argmax works in your own repo and sends changes as normal pull requests. The notes live next to the code, so a copy of the repo holds both.

Where does it run?

On your own servers, with your own AI account. Your code goes only to the AI provider you pick. Every try runs on a separate copy, and your main code only changes when you merge a win.

What if nothing improves?

Then nothing changes. A failed try throws away its code and keeps its notes, so you learn what doesn't work and the next try starts from there. Every change you get has passed its check.

Does this replace Copilot or other coding tools?

No. Argmax needs a number to improve, test problems, a build that runs itself, and a fair comparison. Without those it has nothing to aim at. Use it next to your current tools, on the part of your code where results are the product.

Can my team still steer?

Yes. Add ideas yourself, change the order, and review every change before it merges. Argmax suggests ideas too, but the list is plain GitHub tickets that you control.

What does it cost?

Pricing isn't published yet. We scope it with you on the demo call.

Your numbers, your code

Bring a number. See how to move it.

We’ll check your code against the four needs and walk through a real try, start to finish.

01your number and your test problems
02a quick check against the four needs
03a live try: idea → research → pull request
Book a demo →

Nothing reaches you unless it is clearly better, and nothing merges without your review. A failed try never touches your code: you keep what was learned, not the risk.

Prefer to look first? Request trial access.