Hire AI and machine-learning engineers
Engineers who put models into production and measure them, including retrieval systems over your own documents.
What a Tankar AI engineer does
Builds the evaluation set before the model, ships the system behind an API with monitoring, and reports accuracy, latency and cost per query rather than a demo that works on three examples.
Retrieval first, fine-tuning later
Most document and support problems are retrieval problems. Getting the right passages in front of a capable model, with citations back to the source, beats fine-tuning on cost, effort and the ability to explain an answer.
When we say no
If a rule, a lookup or a small classifier solves the task reliably, we will tell you that instead of selling a model. A system nobody can explain to an auditor is a liability in a regulated process.
What each level brings
| Level | What they bring |
|---|---|
| Junior |
|
| Mid-level |
|
| Senior |
|
Engagement terms
The terms that hold for every developer placed on this stack.
| Term | Detail |
|---|---|
| IP and source code | Full IP assignment at final payment. Source code, prompts and evaluation sets sit in your repository from week one, and an NDA is signed before discovery. |
Related case studies
Platforms built with AI and machine learning developers on the team.
Own productGovernment and public sectorTenderBazaar: tender discovery and alerts for Indian SMEs
A tender discovery platform that collects listings from central, state and municipal portals into one searchable feed with daily alerts.Services: SaaS product development, Mobile app development and 1 moreRead the case study
Anonymised clientRetail and e-commerceA multi-vendor marketplace with vendor self-service and compliant invoicing
A marketplace where vendors onboard themselves, list and price their own stock, and are paid out against a commission engine with GST-compliant invoices.Services: E-commerce and marketplace development, Custom software development and 1 moreRead the case study
Questions buyers ask
What buyers ask before they add a developer on this stack.
Do we need fine-tuning?
Usually not first. Retrieval over your own documents, better prompts and clear evaluation solve most cases at lower cost. Fine-tuning is worth it for a fixed format or a narrow task with plenty of examples.
How do you measure whether the system is good enough?
With an evaluation set built from your real documents and agreed before the build, scored on accuracy for your task, plus latency and cost per query. The scores go in the report, including where the system fails.
What does this cost to run each month?
That depends on the model, the query volume and how much context each query carries. The estimate includes a cost per query and a projection at your expected volume so there is no surprise.
Can they work alongside our data team?
Yes, and it is the better arrangement. Your team knows the data and its history; the engineer brings the evaluation, deployment and monitoring practice.
Request a developer
Tell us the level, how many people and when they should start. A named person replies within one business day. NDA on request.
Other stacks
Every stack has the same page: levels, vetting, terms and a request form.
- Next.js and ReactNext.js, React, TypeScript
- Node.jsNode.js, TypeScript, Express and Fastify
- PythonPython, FastAPI and Django, pandas and Polars
- .NET.NET 8 and later, C#, ASP.NET Core
- FlutterFlutter, Dart, Riverpod and Bloc
- React NativeReact Native, TypeScript, Expo
- DevOps and platform engineeringDocker, Kubernetes, Terraform
- QA and test automationPlaywright, Cypress, Appium
Tell us what you are building.
NDA on request. Written estimate within 48 hours of a scoped call. Reply within one business day.