We built AI for German aerospace. Now we build engineers in Jaipur.
Physical AI apprenticeships on real manufacturing problems — led by an engineer from DLR, Fraunhofer & Tier-1 automotive R&D. Not a course. A lab.
DLR · Fraunhofer · Tier-1 Automotive R&D · M.Sc. (Germany)
Who are you?
I'm an engineering student
I want a career AI can't replace.
- Real industrial problem statements, not toy datasets
- Direct mentorship — DLR/Fraunhofer methodology
- A deployed portfolio project + reference letter
I represent a college
I want an Industry CoE on my campus.
- Industry Center of Excellence at zero capital cost
- Documented industry linkage for NAAC Criterion 3
- From first meeting to running lab in under 30 days
I represent a company
I have an engineering problem worth solving.
- Defence, automotive & manufacturing, pharma
- Simulation surrogates, design & process optimisation
- Starts with a scoped feasibility study on your own data
Certificates are dead. Recruiters know it. You know it.
India produces 15 lakh engineers a year with the same certificates. Interviews are won by the 1% who can demo a deployed system and explain every line.
| Criterion | Certificate course | Karkhana Lab Industry-Incubator |
|---|---|---|
| What you build | What you buildQuizzes and toy notebook exercises | A deployed industrial system |
| Who teaches | Who teachesPre-recorded videos | An engineer from DLR / Fraunhofer / Tier-1 R&D |
| What you show recruiters | What you show recruitersA PDF certificate | A live demo you can defend line by line |
| Class size | Class sizeUnlimited enrollment | 25 seats, screened-in |
| What it proves | What it provesYou paid for a course | You can ship production work |
This is what our lab actually builds.
Satellite Telemetry Anomaly Detection
German Aerospace Center (DLR), Köln · 2021
Catches failure precursors in satellite telemetry — and refuses to trust its own alarm when the evidence is contaminated.
View case
BEWiS — Weightless Neural Networks on GPU
Fraunhofer FKIE, Wachtberg · 2019–2020
Real-time background subtraction that learns what "normal" looks like from the video itself — no training set, no labels.
View case
Physical AI for Design & Material
Applied R&D with a global Tier-1 automotive manufacturer · 2022–2026
CAD goes in, a graph neural network learns the part’s physics, and the shape itself becomes something an optimiser can search — in minutes.
View case
Physical AI for Process
Applied R&D with a global Tier-1 automotive manufacturer · 2022–2026
You bring the CAD. The geometry is never assumed — the process window is optimised for the part you actually upload.
View case
Blow-Mould Wall Thickness — FEM in Seconds
Applied R&D with a global Tier-1 automotive manufacturer · 2022–2026
A neural network reproduces a 45-minute finite-element wall-thickness simulation in seconds — and the comparison shows where it still disagrees.
View case
Optimisation
Lab teaching track · used across the design, material and process work
Every project on this page ends in the same question: of all the settings you could choose, which one do you actually run? This is how that gets answered.
View case
Land Decision Engine
Independent build · 2025–2026
Turns "should I hold, build on, or sell this plot — and when?" into a scored, auditable answer.
View case
Market Research Dashboard
Industry collaboration · 2025–2026
A market analysis platform built to be honest about what it cannot know — and audited until it was.
View case
Learn from someone who has actually built it.
M.Sc. Autonomous Systems, Bonn-Aachen International Center for Information Technology (b-it) · anomaly detection & Explainable AI for satellite operations at DLR · weightless neural networks on GPU at Fraunhofer FKIE · established and led the AI team for global OEM projects at a Tier-1 automotive manufacturer.