Job Description
A typical day in office :
- You are developing video perception algorithms for unsolved problems in the technology area perception for autonomous drive vehicles.
- You have to explain why something works or not using data. And what next for the things not working with the reduced cost of labelling.
- You have to find that elusive data through innovative algorithms,
Your job would involve
Looking at the nature of the dataCurrent AI / DL algorithm performanceDerive insights or patterns in failuresDefine a hypothesis for the solution making.Define Experiments for hypothesis testingDefine / Review the techniques / methods in deep learning for experimentsAnalyze the results of experimentsDefine the next set of data or experiments forwardDefine the targets for the current sprintReview, Agree on the experimentsEnsure delivery with the right process for qualityYou are a skilled software engineer with experience in C++ and strong problem-solving skills.You are passionate about solving real-world robotics problems, and you have ideally worked on autonomous robots before. Prior knowledge of working with deep neural networks is a strong plus.A team player.You take ownership and work with the team to deliver exceptional results.You are interested in the performance of the entire system across engineering disciplines.Ability to build and iterate quickly (AGILE mindset). You enjoy working fast and smart, and you are comfortable in the earlier stages of developing an algorithm from scratch.Hands on. You dig deep into important details such as the sensor driver if it improves the overall system. You like working with production machine learning pipelines, from dataset collection and labeling to training and validation.Great communicator. You have experience writing clear, concise, and detailed documentation.Should be able to explain technical solutions clearly to non-technical / different audienceProblem solver : You are expected to solve systemic problems.Skills : Must have : 1. Strong knowledge in DL (3yrs min)(few-shot learning, transfer learning, unsupervised and semi-supervised methods, active learning and semi-automated data annotation, large scale image and video detection and recognition)2. Hands on experience in Full stack Vision / Image related projects in production (3yrs min)
Data : Spark, Hadoop (user level), visualization, data preparation, data handling, understanding data
Model : tensor flow 2+ and PyTorch
Deployment : docker, flask, kubernets
Cloud deployment : Azure / AWS3. Experience in Python4. Tool :
GiT
Working with GPU
Linux
Good to have : Coding :
Multi-processing and multi-threading (application)
Understanding of and C++ (for development and deployment)
Sound knowledge of custom accelerator deployment (OpenVINO, Jetson)
Tooling :
Good understanding of Visual studio code
CUDA, RAPIDS, OpenMP, Distributed Systems and HPC application designSkills Required
Linux, Gpu