Short DescriptionShould be able to independently work with Deep Learning models related to computer vision like CNNs, generative models etc. Maintain a good overview understanding of latest architecture models (E.g. GANs, very deep / wide / dense / gated / skip models), algorithm trends (e.g. Adam) and techniques (e.g. for regularization or initialization) in deep-learning for computer-vision and judge their relevance to solve enterprise use-cases.
- Doctoral, Master's or Bachelor's degree in Engineering or Technology, majoring in Computer Science, from a reputed institute, with excellent academic records.
- Hands-on experience with at least one major deeplearning library for computer-vision like TensorFlow Torch or Theano.
- Should be proficient in python and specifically in libraries used for machine-learning / computer-vision like openCV, numpy, scikit-learn, keras, matplotlib and scipy.
- Ability to decipher the mathematical basis of machinelearning / deep-learning models based on linear-algebra, probability & statistics, vector / tensor algebra and calculus will be an added advantage, but is not mandatory.
- Hands-on experience with scalable & parallelizable scripting algorithms that cater to large-scale image and video datasets used in conjunction with deep-learning for computer-vision.
- Strong analytical and problem solving skills.
- Goal-oriented team player with flexible mindset
- Good communication skills, fluent in English, both written and spoken.
- Should be able to independently work with Deep Learning models related to computer vision like CNNs, generative models etc.
- Maintain a good overview understanding of latest architecture models (E.g. GANs, very deep / wide / dense / gated / skip models), algorithm trends (e.g. Adam) and techniques (e.g. for regularization or initialization) in deep-learning for computer-vision and judge their relevance to solve enterprise use-cases.
- Should be able to rapidly develop, deploy, test and iterate with such architectural models / algorithms based on both inhouse and open-source research (concept-to-code) and guide the team to select the best approach for productive deployments.
- Own the training life-cycle of deep-leaning based computervision models, including data-set preprocessing, augmentation, annotation tools and techniques, rapid iterations for hyper-parameter setting, interpretation of test result matrices and plots, fine-tuning and optimization.
- Maintain an overview understanding of a deep learning technology stack from GPU based H/W through cuda / cudnn to framework implementation (in c++ or python) and use this know-how for optimizing the training and prediction workloads, dependency handling and performance tuning.
- Brings hands-on flavor to all aspects in software engineering related to enterprise grade deep-learning.
- Should be a good learner who takes initiative to drive topics to completion with a high quality, reliability and efficiency.
- 4 - 8 years professional experience with at least 2 years of hands-on experience in deep learning, preferably with computer vision in an enterprise product development setting, in the role of a data scientist / deep learning senior engineer.
- Experience in working with 3rd party tools and open source libraries relevant for the field of computer vision.
- Contribution to open source community, participation in competitions (like Kaggle), certifications, paper publishing and presentations in relevant journals / conferences / events / meet-ups will be considered as additional merit.
- Prior experience in working with Agile software methodologies (Scrum) and iterative development cycles for deep learning
Information Systems Engineer