Job Location: Delhi
Description:
Roles and Responsibilities:
Analyze and Improve the Existing Data, Training & Inference Architecture for existing Use-Cases.
Pioneer Novel Algorithms for new use-cases to develop an elegant and robust solution to solve problem statements using Computer Vision.
Work on Model Optimization to improve Performance & Throughput for the same.
Work on Various SOTA Dataset Augmentation & Dataset Generation Techniques that help improve model accuracy & precision.
Work on development & improvement of End-to-End Pipeline use-cases running at scale.
Work with the existing teams to set up a re-training pipeline that iteratively improves existing model accuracy based on in-coming datasets.
Requirements:
Proficient with Training of Detection/Classification/Segmentation Models with Tensorflow/PyTorch
Good Theoretical and Practical Knowledge with the fundamentals of Deep Learning, eg. CNNs, Regularization Techniques, etc.
Proficient in developing Deep Learning Model Pipelines
Deployment of Deep Learning Models at Scale for live processing of video feeds (RTSP).
Experience with Model Servers: Triton Inference Server (Preferred), TensorFlow Serving, or Torch Serve.
Experience in working with Model Optimization for Nvidia Hardware (TensorRT Conversion of both Tensorflow & Pytorch models)
Experience with both Cloud as well as Edge Deployment of Deep Learning Models (Preferably at Scale).
Experience in FP16/Mixed Precision and INT8 Optimized Inference on Nvidia Hardware
Experience with using Docker, Nvidia-Docker, Nvidia NGC containers for Computer Vision/Deep Learning.
Proficient in Data Structures & Algorithms
Proficient understanding of code versioning tools, such as (Git / Mercurial / SVN)
Experience with common front-end development tools such as Babel, Webpack, NPM, etc.
Knowledge of a state management library such as Redux, etc
Note: CTC will be based on seniority.
Good to have
Experience with Scalable Cloud Deployment Architecture for Video Analytics (Involving Kubernetes and/or Kafka)
Experience in INT8 Training (Quantization Aware Training)
Experience/Knowledge of TensorFloat32 and BFloat16 Training/Inference
Note : CTC will be based on Seniority
How we work
1. Microsoft Teams is our go to tool for daily communication
2. We conduct daily standups and team meetings over Teams
3. We encourage open discussion and debate, and place tremendous value on ownership and founder mindset.
4. We value: design + amazing UI/UX, documentation, To-Do lists and Data based decisions
5. We’re distributed but not alone! We might be working around the world but we believe in the value in getting to know each other outside of working hours
6. We get together for a bi-weekly townhall and frequent destressing session with a certified healer!
Perks and Benefits :
1. 100% remote workplace model
2. Health Insurance of Rs. 3L
3. Top performers eligible for attractive equity options
4. Mental Health Consultations with heyy app.
5. Company sponsored upskilling courses
6. Growth Hours: 40 hours every year
7. The chance to give back with 40 hours for community causes
8. Fun company retreats twice a year
9. Access to a Financial Advisor
Submit CV To All Data Science Job Consultants Across India For Free
Leave a Reply