AI feature sprint
For founders and teams
You have a product and one clear job for AI: answering questions from your documents, drafting replies, or running a workflow that calls your APIs. I build that feature into your product and test it before your users see it.
Many teams ship AI features without testing them properly. Every sprint ends with an eval set, so you can see where the feature is right, where it isn't, and what it costs to run.
Based in Bengaluru, India. I meet clients in person in Bengaluru and work remotely with clients worldwide.
- Starting price
- from ₹75,000 · $1,500
- Timeline
- 1 to 2 weeks
How it works
Free call
We talk through the feature, your data and what it must never do.
Written scope
A short scope with the first eval questions, so we agree on what good looks like before I build.
Build
I build the feature into your product, with a short update every day.
Eval report and handover
The eval results, docs on how it works and how to change it, and a handover call.
Two weeks of fixes
Anything that comes up after delivery, fixed at no charge.
What you get
- The feature, built into your product
- An eval set and a report of how it scores
- Guardrails: refusals, rate limits, cost caps and safe fallbacks
- Handover docs and a walkthrough call
- Two weeks of fixes after delivery
Questions
- What's an eval set?
- A list of real questions with the answers you expect. It runs against the feature, so you can see where it's right and where it isn't before your users do, and run it again after every change.
- Which AI models do you work with?
- OpenAI and Gemini, or the provider you already use. It runs on your own account and keys.
- Who pays for the AI usage?
- You do, on your own API keys. I add cost caps and rate limits so a busy day can't run up the bill.