About the role
In this role, you will take end-to-end ownership of converting complex problem statements into scalable, state-of-the-art machine learning solutions. Working at the intersection of data engineering, research, and deployment, you will harness LLMs, VLMs, and Transformer architectures to build, optimize, and version high-impact models alongside a collaborative team of ML engineers.Responsibilities
- Handle all aspects of data filtering and agglomeration.
- Collaborate with other ML Engineers to brainstorm on best practices, updates on SOTA models.
- Handle all translation of problem requirements to choice of architecture, codebase, and datasets. Enhancements of in-house/available open-source repositories to meet problem requirements.
- Work with data labeling teams to handle annotation metadata for your datasets.
- Handle all aspects of the Machine Learning cycle including data preprocessing, augmentations, training, model and experiment versioning, ablation studies and benchmarking.
- Bachelors or Masters Degree in Computer Science or related disciplines.
- 2 - 4 years of experience required.
- At least 1 year of experience with LLMs/VLMs including their architecture, training, datasets, and knowledge of SOTA models and frameworks to leverage these models.
- Strong theoretical and prior experience of CNNs and Transformers including their usage in at least one successful product deployment
- Fluency in Python Programming.
- Working knowledge of C/C++ desirable.
- Familiarity with ML Frameworks like Pytorch, TensorFlow etc.
- Familiarity with MLOps, including Model Quantization, usage of annotation tools and model versioning tools like MLFlow.
- Strong background in Mathematics and Statistics.
- Sharp problem solving skills and ability to resolve ambiguous requirements
- Familiarity with Dockers.