/ faq
Questions, answered plainly.
The things people ask us most. If yours is not here, write to us.
/ company
About AlgoQX.
What does AlgoQX do?
We help teams test and improve language models and agents: evaluation, post-training with human feedback (RLHF) and verifiable rewards (RLVR), and the verifiers and environments behind them.
Who do you work with?
Teams shipping agents, startups building on language models, research groups, and companies adapting open models. See who we work with.
Do you pretrain foundation models?
No. Our focus is post-training and evaluation of existing models.
Do you sell a product?
Not yet. We work through engagements today. When a product is ready, it will be announced here with its real status.
/ working together
Engagements.
Can you test an agent we already built?
Yes. That is the most common starting point: an evaluation on tasks that reflect how your users actually use it.
Can you train a model for our task?
Yes, on open-weight models, with evaluation built in so you can see whether training helped.
Which models do you work with?
Mostly open-weight models. We discuss what fits your task, constraints and infrastructure during scoping.
How do you handle our data?
Data access is agreed during scoping and kept to what the work needs. Please do not send confidential data in a first message.
How is it priced?
Per project, agreed before work starts. See pricing.
How do we start?
Email sales@algoqx.tech with a short description of the task and the model.