Skip to content

Hire · Freelance & consulting

Hire a TensorRT developer to make your models fast enough for the edge

When a model has to run on a Jetson instead of a GPU server, I export it through ONNX to TensorRT, benchmark it on your hardware, and report what you actually get — not what the marketing page promises.

01

Problems I solve with this stack

The model runs fine on a workstation but can't keep up on the edge device.

Export to ONNX/TensorRT with the right precision (FP16/INT8 where acceptable) and benchmark — this is the named next step for the CCTV build to get back to higher frame rates.

No one knows what latency to promise the client.

A measured throughput report on your hardware: streams, frame rate and latency you can expect.

Edge devices are hard to update.

Containerised deployment and a repeatable build for model artefacts, so updates are a swap, not a surgery.

02

What you get

  • ONNX export and TensorRT engine build for your model.
  • Precision tuning (FP16/INT8) with an accuracy-vs-speed comparison.
  • Edge deployment on NVIDIA Jetson (Docker, systemd, or both).
  • Benchmark report on your hardware.
  • Handover: how to rebuild the engine when the model changes.
03

Frequently asked questions

Do you work with international clients?

Yes. I'm based in Pune, India, and work with teams in India and remotely worldwide. Communication is async-first with scheduled calls when needed.

What does an engagement look like?

A short call to scope the problem, a written scope (what will be built, what won't, and how we'll know it works), then small milestones you can review — with the code in your repository from day one.

Can you work with our existing codebase?

Yes. Most work is hardening, extending or extracting from existing systems rather than greenfield. I'll flag honestly where a rewrite isn't worth it.

Are you available for full-time roles too?

Yes — freelance, consulting and full-time conversations are all welcome. See the work-with-me page for how I work with clients and employers.

More questions? Ask directly

Need a tensorrt build?

Book a free 15-minute call below — describe the problem, I'll tell you honestly whether I'm the right person for it.

How I work with clients