👋 Welcome
I’m Dinesh C. Dhotrad, a Computer-Vision & Robotics engineer who loves turning cutting-edge research into real-time, production-ready systems.
Recently completed my M.S. in Computer Science (Case Western Reserve University, ’25) and am now open to full-time roles in AI, 3-D CV, and Robotics.
🔭 What I’m Building
| Project | Highlights |
|---|---|
| Decentralized 3-D Generation (GPU.net, 2023-24) | NeRF + photogrammetry across 5 000+ volunteer GPUs → 30 % cloud-cost reduction |
| 360° Video-to-3-D Pipeline (Tooliqa, 2021-23) | Proprietary 3-D voting ↓ noise 90 %, UV texturing 15× faster |
| Multi-GPU LLM/Vision Training | GPipe pipeline parallelism → 42 % throughput boost on A100 cluster |
| 2-D→3-D Point-Cloud Segmentation | OneFormer masks + back-projection voting, outperforming PointNet baseline |
| RoadSense ADAS | YOLOv8 + classical vision for lane & object detection on Jetson |
🛠 Core Expertise
- 3-D Computer Vision / SLAM • LiDAR/RGB-D fusion, multi-view geometry
- Deep Learning & HPC • PyTorch, TensorRT, CUDA, NCCL, mixed-precision training
- Embedded & Edge AI • NVIDIA Jetson/Orin, ROS 1, real-time optimization
- Algorithms & Math • Kalman/particle filters, linear algebra, signal processing
📄 Curriculum Vitae
Professional Summary
Results-driven Computer-Vision & AI Engineer with 3 + years’ experience developing and optimizing deep-learning and HPC pipelines for autonomous systems. Proven track record in GPU acceleration, photogrammetry, and real-time deployment on Jetson and cloud platforms.
Technical Skills
Languages: Python, C++, C, CUDA
Frameworks: PyTorch, TensorRT, Scikit-Learn, Keras
CV/3-D: OpenCV, Open3D, SLAM, NeRF, GANs
HPC: CUDA, OpenMP, multi-GPU training, distributed computing
Cloud & DevOps: AWS SageMaker, Docker, Kubernetes
Embedded: NVIDIA Jetson, Raspberry Pi + NPU, ROS 1
Experience
3-D AI Lead Engineer — GPU.net (Bengaluru)
Jun 2023 – Feb 2024
- Parallelized text/image/video-to-3-D pipelines (NeRF + photogrammetry).
- Designed decentralized GPU framework → 30 % cloud-cost savings.
- Containerized privacy-preserving training workloads with Docker.
Software Engineer – Computer Vision — Tooliqa (Gurugram)
Mar 2021 – Apr 2023
- Built 360° video-to-3-D pipeline; proprietary voting algorithm cut noise 90 %.
- Improved LiDAR/IMU fusion accuracy 25 % for indoor reconstruction.
- Achieved 15× speed-up in SLAM texturing via optimized C++ projection code.
Research Intern — KLE Technological University
Aug 2020 – Feb 2021
- GAN-based underwater image enhancement; visibility ↑ 60-75 %.
Project Trainee — IIT Delhi / AIIMS
Jun 2019 – Jul 2019
- Assisted in AI-driven surgical micro-suturing assessment tool.
Education
| Degree | Institution | Year |
|---|---|---|
| M.S. Computer Science | Case Western Reserve University | 2025 |
| B.E. Computer Science | KLE Technological University | 2021 |
Key Projects
- Multi-GPU Pipeline Parallelism for LLMs – 42 % training speed-up.
- Point-Cloud Segmentation via 2-D Masks – hybrid approach, reduced compute.
- Real-Time RoadSense ADAS – YOLOv8 + classical CV on Jetson.
- Adversarial ML Security – exposed classifier weaknesses via crafted perturbations.
Certifications
NVIDIA CUDA C/C++, IIIT-H 3-D Vision, Coursera DL Specializations, AWS Cloud-Native, NPTEL C Programming.
Additional
Languages: English, Kannada, Hindi
Interests: Open-source AI, autonomous vehicles, photography, motorcycle touring
(Download full PDF résumé → Dinesh_Dhotrad_Resume.pdf)
🌱 Notes Garden Roadmap
- Project deep dives (architecture diagrams, benchmarks)
- Research reviews (Gaussian Splatting, Diffusion for 3-D)
- HPC playbook (MPI, Slurm scripts, Triton kernels)
- Robotics logs (ROS launch configs, sensor-fusion experiments)
Stay tuned as this graph grows!
📫 Connect
LinkedIn • dineshdhotrad99@gmail.com • +1 (972) 363-3446
Always eager to discuss perception stacks, HPC tricks, or cool robotics ideas.