Experience
May, 2023 - Present
Research Assistant
TAmpere Software Engineering (TASE), Tampere University
LiquidAI Project (Jan, 2024 - Present)
I achieved performance improvements for machine vision tasks:
- Benchmarked deep learning models (YOLOv5, MobileNetV2) for images on edge devices: NVIDIA
Jetson Orin NX, Raspberry Pi 3 & 4, and Intel NUC.
- With GPU utilization strategies and CUDA context management, I improved YOLOv5 image
inference time to ≈15ms/image using TensorRT, ≈40% and ≈97% improvement compared to Intel
NUC and Pi 3 respectively.
- Analyzed effect of library import times, model/image loading, warmup, inference, and
post-processing on Jetson device’s performance on 1000 iterations.
Cloud Technologies (May-Dec, 2023)
Tampere University Masters course where I developed exercises with instructor in May-Nov 2023. I
built exercises on VMs, Cloud Computing, Docker, Terraform, Ansible, networking basics etc. on
OpenStack based IaaS cloud infrastructure on top of cPouta.
DevOps course support
Setting up self-hosted Gitlab server and runner for students to submit projects with CI/CD for
COMP.SE.140-DevOps course on the IaaS service cPouta.
Things I Know
Languages:
Python3, C++, Bash, C#, Rust (Wasm)
Tools & Frameworks:
Tensorflow, Pytorch, TensorRT SDK, GPU Programming, Docker, Terraform, Ansible, OpenStack based cloud, VMWare
Version Control:
Git, self-hosted Gitlab
Document Preparation:
Overleaf, Notion
Soft skils:
Team collaboration, problem solving
Also: Computer Vision stuff on GPU devices
