What are the responsibilities and job description for the Machine Learning Engineer/AI Engineer position at eTeam?
Job Title: Machine Learning Engineer/AI Engineer
Location: Open To Hubs in Atlanta GA 30328 and Minneapolis MN 55402
Hybrid: 60% In Office Required (3 Days Per Week)
Working Hours: 8am - 5pm CST
Duration: 8 months on contract with possible extension
TOP REQUIREMENTS:
• Fiddler tool (3-5 years)
• Lumenova tool (3-5 years)
• Azure/cloud knowledge (3-5 years)
• AI use case implementation (3-5 years)
JOB DESCRIPTION:
To effectively monitor using Fiddler AI or Lumenova tools implemented use cases, key skills include:
• understanding AI model metrics, data analysis, statistical knowledge, familiarity with bias detection techniques, prompt engineering, LLM specific monitoring capabilities, visualization skills, and the ability to interpret model explanations to identify potential issues like hallucinations, safety violations, and data drift; all while being able to leverage Fiddler AI's or Lumenova features for comprehensive monitoring and analysis across different AI applications
Specific skills to focus on when monitoring Fiddler AI and Lumenova AI tools:
• Model performance metrics:
1. Accuracy, precision, recall, ,-score, perplexity, BLEU score (depending on the AI task)
2. Monitoring changes in these metrics over time to detect model drift
• Data analysis:
1. Identifying patterns and anomalies in input data and model outputs
2. Analyzing data distributions to detect potential biases
• Bias detection:
1. Understanding protected attributes and how to assess bias across different demographic groups
2. Utilizing Fiddler's fairness metrics like disparate impact and equal opportunity
• Prompt engineering:
1. Crafting effective prompts to test model behavior and identify potential issues
2. Understanding how different prompts can influence model responses
• LLM specific monitoring:
1. Identifying and mitigating hallucinations (generated text that is factually incorrect)
2. Detecting safety violations like toxic language or personal information leaks
3. Monitoring for prompt injection attacks
• Visualization skills:
1. Creating dashboards and visualizations to effectively communicate model performance and potential issues
2. Using tools like 3D UMAP to identify clusters and patterns in data
• Model explainability:
1. Interpreting model explanations provided by Fiddler to understand why a model made a certain prediction
2. Utilizing these explanations to identify areas for model improvement
• Fiddler AI or Lumenova platform knowledge:
1. Familiarity with Fiddler's / Lumenova features like data ingestion, model monitoring, bias detection, and root cause analysis
• Azure/cloud knowledge: maintain/debug the cloud platform stack working with the cloud engineers, work with vendors to upgrade the software, debug user issues
ET_RV01
Location: Open To Hubs in Atlanta GA 30328 and Minneapolis MN 55402
Hybrid: 60% In Office Required (3 Days Per Week)
Working Hours: 8am - 5pm CST
Duration: 8 months on contract with possible extension
TOP REQUIREMENTS:
• Fiddler tool (3-5 years)
• Lumenova tool (3-5 years)
• Azure/cloud knowledge (3-5 years)
• AI use case implementation (3-5 years)
JOB DESCRIPTION:
To effectively monitor using Fiddler AI or Lumenova tools implemented use cases, key skills include:
• understanding AI model metrics, data analysis, statistical knowledge, familiarity with bias detection techniques, prompt engineering, LLM specific monitoring capabilities, visualization skills, and the ability to interpret model explanations to identify potential issues like hallucinations, safety violations, and data drift; all while being able to leverage Fiddler AI's or Lumenova features for comprehensive monitoring and analysis across different AI applications
Specific skills to focus on when monitoring Fiddler AI and Lumenova AI tools:
• Model performance metrics:
1. Accuracy, precision, recall, ,-score, perplexity, BLEU score (depending on the AI task)
2. Monitoring changes in these metrics over time to detect model drift
• Data analysis:
1. Identifying patterns and anomalies in input data and model outputs
2. Analyzing data distributions to detect potential biases
• Bias detection:
1. Understanding protected attributes and how to assess bias across different demographic groups
2. Utilizing Fiddler's fairness metrics like disparate impact and equal opportunity
• Prompt engineering:
1. Crafting effective prompts to test model behavior and identify potential issues
2. Understanding how different prompts can influence model responses
• LLM specific monitoring:
1. Identifying and mitigating hallucinations (generated text that is factually incorrect)
2. Detecting safety violations like toxic language or personal information leaks
3. Monitoring for prompt injection attacks
• Visualization skills:
1. Creating dashboards and visualizations to effectively communicate model performance and potential issues
2. Using tools like 3D UMAP to identify clusters and patterns in data
• Model explainability:
1. Interpreting model explanations provided by Fiddler to understand why a model made a certain prediction
2. Utilizing these explanations to identify areas for model improvement
• Fiddler AI or Lumenova platform knowledge:
1. Familiarity with Fiddler's / Lumenova features like data ingestion, model monitoring, bias detection, and root cause analysis
• Azure/cloud knowledge: maintain/debug the cloud platform stack working with the cloud engineers, work with vendors to upgrade the software, debug user issues
ET_RV01
Salary : $69 - $81
