Required Skill Sets on top of the above skills:
Role details
Job location
Tech stack
Job description
- Knowledge on deploying models on mobile devices iOS/Android
- Knowledge on C++ for custom functions and writing unit test cases.
- Strong debugging skills on C++/Python code.
Basic jargons of ML which include Cost functions, Gradient Descent, Back Propagation, Activation functions etc
Requirements
-
Experience in Data Science and DeepLearning frameworks.
-
Customer requirement analysis, cross team collaboration
-
Software Development Lifecycle, strong Software Design/Development experience
-
Computer Science or Computer Engineering or equivalent technical degree
-
must be able to recognize potential issues, and compose technical communications in GitHub)
-
Experience working with Windows, MacOS, and Ubuntu environments
-
Excellent written and oral communication skills
-
Being a team player with a positive attitude and people skills
-
Open to learning new internal technical tools, * Python installation, environment setup and Jupyter Notebook
-
Object and Data Structures basics
-
Comparison Operators and Statements
-
Methods and Functions
-
Errors and Exception handling
-
Built-in functions and Python Generators
-
Using scientific Python libraries numpy, pandas, matplotlib, scikit-learn
-
Use data visualization with Python, * Multi-Backend Installation: Installing Keras 3 and configuring backends (JAX, PyTorch, or TensorFlow) using the KERAS_BACKEND environment variable.
-
Core Data Structures: Understanding Layers, Models, and the fundamental difference between the Sequential API, Functional API, and Model Subclassing.
-
Backend-Agnostic Ops: Familiarity with the keras.ops namespace (the cross-framework NumPy-like API) and keras.random for writing framework-independent code.
-
State Management: Concepts of statelessness vs. statefulness, especially when working with the JAX backend and Keras 3's functional layer calls.
-
Training & Evaluation: Mastering the high-level .fit(), .evaluate(), and .predict() workflows, as well as writing Custom Training Loops using GradientTape (TF/PyTorch) or jax.grad.
-
The Distribution API: Knowledge of keras.distribution for multi-GPU and TPU training (Data Parallelism and Model Parallelism).
-
Optimization & Compilation: Understanding XLA (Accelerated Linear Algebra) and how to leverage jit_compile for performance across different hardware.
-
Serialization: Using the modern .keras v3 format for saving/loading models across different frameworks and platforms.