Next Levels
Lead Machine Learning Engineer
Next Levels, Jackson, Mississippi, United States,
A Series A Company arerevolutionizing the shopping experience using the power of generative AI and rich messaging technologies to build a personalized shopping assistant for every consumer.
The Role
They are hiring ML Engineers, preferably with a background in NLP, to work on all aspects of dialog systems, from fine-tuning LLMs to serving models and building data pipelines and entity resolutions. If you don’t have an NLP background but have deep experience with Python and some exposure to ML and you are excited about the space and willing to learn, we would still encourage you to apply.
Key Responsibilities:
Build and maintain a training pipeline for fine-tuning open-source LLMs. Testing different LLM architectures at different sizes and quantizations is important.
Serve and Deploy the LLM on GPUs. It is important to know different configurations ahead of time to ensure we don’t OOM at runtime.
Formulate novel product issues to ML problems and propose and iterate on solutions quickly. We value the fast pace of development.
Triage customer issues to model/system shortcomings and communicate the complex architecture in a simple, comprehensible way to internal stakeholders.
Keep up with the fast-growing generative AI space and incorporate the latest developments into our system.
You
MS/Ph.D. in computer science, mathematics, or another quantitative discipline or exceptional work experience in ML engineering.
5+ years ML Experience.
Experience developing machine learning-driven products at scale.
Experience solving problems using Machine Learning with PyTorch.
Exceptional understanding of designing ML systems in production at scale and applying a quantitative approach to identifying performance bottlenecks and cost/performance tradeoffs.
Practical experience with large data systems, data models, and batch and streaming data pipelines.
Ability to ramp up quickly on our tech stack, which includes GCP, Kubernetes, Airflow, Pandas, PyTorch, Python, and Node.js.
Extensive experience with model deployment, both static within applications and dynamic, using Flask, FastAPI, or similar frameworks.
#J-18808-Ljbffr
The Role
They are hiring ML Engineers, preferably with a background in NLP, to work on all aspects of dialog systems, from fine-tuning LLMs to serving models and building data pipelines and entity resolutions. If you don’t have an NLP background but have deep experience with Python and some exposure to ML and you are excited about the space and willing to learn, we would still encourage you to apply.
Key Responsibilities:
Build and maintain a training pipeline for fine-tuning open-source LLMs. Testing different LLM architectures at different sizes and quantizations is important.
Serve and Deploy the LLM on GPUs. It is important to know different configurations ahead of time to ensure we don’t OOM at runtime.
Formulate novel product issues to ML problems and propose and iterate on solutions quickly. We value the fast pace of development.
Triage customer issues to model/system shortcomings and communicate the complex architecture in a simple, comprehensible way to internal stakeholders.
Keep up with the fast-growing generative AI space and incorporate the latest developments into our system.
You
MS/Ph.D. in computer science, mathematics, or another quantitative discipline or exceptional work experience in ML engineering.
5+ years ML Experience.
Experience developing machine learning-driven products at scale.
Experience solving problems using Machine Learning with PyTorch.
Exceptional understanding of designing ML systems in production at scale and applying a quantitative approach to identifying performance bottlenecks and cost/performance tradeoffs.
Practical experience with large data systems, data models, and batch and streaming data pipelines.
Ability to ramp up quickly on our tech stack, which includes GCP, Kubernetes, Airflow, Pandas, PyTorch, Python, and Node.js.
Extensive experience with model deployment, both static within applications and dynamic, using Flask, FastAPI, or similar frameworks.
#J-18808-Ljbffr