A small studio for ML systems that need to ship.
spacedrift is a Bengaluru ML & AI studio run by Lourdu Raju, a Machine Learning Engineer. We take on fixed-scope work for teams that need practical output, not a long agency engagement.
Why it exists
Researchers need experiments they can reproduce. Startups need AI that works on their own data. ML teams need labels they can audit. These projects fail when the scope is loose and nobody owns the build.
So the studio is small on purpose. spacedrift is a registered MSME run alongside a full-time engineering role. That means fewer projects, each one defined before work starts. Fixed scope and fixed price protect both quality and availability.
You work directly with the person building your system, so requirements, tradeoffs, and delivery questions stay close to the code. If a request really needs a retainer or a large team, we say so up front instead of pretending otherwise.
Principles
What drives every decision.
Quality over volume
We cap active projects so every engagement gets direct engineering attention. If the timeline, data, or scope isn't realistic, we say no.
Scope in writing
Deliverables, milestones, assumptions, and exclusions are agreed before work starts. If delivery risk changes, you hear about it early.
You own everything
Source code, datasets, assets, and documentation are handed over. We deploy to your accounts and avoid needless recurring dependencies.
Fast, not rushed
Speed comes from tight scope and fewer handoffs, never from skipping review, testing, or documentation.
Direct line
You talk to the person writing the code. Technical decisions get made faster and context doesn't get lost between meetings.
30 days included
Bug fixes, small adjustments, and handoff questions for 30 days after delivery. It's part of the scope, not an upsell.
Let's work together.
Tell us what you're building. You'll get a reply within 24 hours.