Job Title: - Data Scientist
Location: - Auburn Hills, MI (Hybrid)
Duration: - Long Term Contract (W2)
Description:
Experience Level: 8+ years of data science experience
Position Summary
We are seeking a highly technical Data Scientist with deep cloud experience, primarily AWS, to design, build, and operationalize machine learning and AI/LLM solutions. This role requires strong engineering discipline (containers, CI/CD, DevSecOps), current knowledge of generative and agentic AI, front-end delivery of model outputs, and a rigorous approach to governance, security, monitoring, and measurable model quality across the full model lifecycle.
Key Responsibilities
Model Development & AI/ML Engineering
- Design, build, train, and validate machine learning models with deep understanding of the underlying data, feature engineering, and model behavior.
- Develop solutions using LLMs and generative AI, including OpenAI modules/APIs, staying current with the latest AI/LLM model releases and capabilities.
- Design and implement agentic AI solutions (multi-step, tool-using, autonomous/semi-autonomous agents), understanding orchestration, memory, and tool-calling patterns.
- Build and evaluate RAG (Retrieval-Augmented Generation) solutions, including semantic RAG architectures (embeddings, vector search, semantic chunking/retrieval strategies).
- Exercise sound judgment on when to apply AI/LLM solutions vs. traditional deterministic or statistical approaches, and select the appropriate model type/size/architecture for a given problem.
- Design for human-in-the-loop (HITL) and human-on-the-loop (HOTL) patterns appropriately, determining where human review, approval, or oversight is required in inference, retraining, or tuning workflows.
- Establish clear, measurable testing and evaluation criteria for model builds (accuracy, precision/recall, drift, latency, cost, hallucination rate, bias metrics).
- Write and maintain automated test cases for model validation, including using AI-assisted tools to generate and expand test coverage for model builds.
Operational Support & Model Lifecycle
- Provide operational support for deployed models, including monitoring, incident triage, and troubleshooting of production ML/AI services.
- Implement governance and monitoring frameworks around deployed models to track performance, drift, bias, and usage over time.
- Own the model update lifecycle: retraining, fine-tuning, versioning, and periodic re-validation as data and business conditions evolve.
- Use logging/chronicle-based tracing and audit trails to track model decisions, retraining events, and lineage over time.
Cloud, Engineering & Front-End
- Build and deploy models and pipelines primarily on AWS (e.g., SageMaker, Lambda, S3, ECS/EKS, Bedrock); working knowledge of GCP and Azure is a plus.
- Strong coding skills in Python and Java for model services, pipelines, and backend integration.
- Build interactive front-end UI applications to present model outputs and insights using React, TypeScript, or Java-based frameworks.
- Containerize model workloads using Docker/containers, and manage GPU-based compute for training and inference workloads.
- Build and maintain CI/CD pipelines for model training, validation, and deployment.
- Apply DevSecOps principles across the ML lifecycle: security scanning, secrets management, infrastructure as code, and automated compliance checks.
Governance, Security & Responsible AI
- Maintain deep awareness of governance, legal, and security requirements applicable to AI/ML model development and data usage.
- Design and implement guardrails in model development (data privacy, bias mitigation, content safety, access controls, prompt injection defenses for LLM/agentic systems).
- Ensure models and pipelines meet organizational and regulatory compliance requirements prior to production release.
Required Skills & Qualifications (Mandatory)
- Strong hands-on experience building and deploying ML solutions on AWS.
- Proven experience with LLMs, including OpenAI models/APIs, and current knowledge of leading AI/LLM model families.
- Hands-on experience building agentic AI systems (multi-agent orchestration, tool use, autonomous workflows).
- Experience building RAG systems, including semantic RAG (embeddings, vector databases, semantic retrieval).
- Deep understanding of data: exploration, quality, feature engineering, and its impact on model outcomes.
- Strong coding proficiency in Python and Java.
- Experience building front-end interactive applications (React, TypeScript, or Java-based UI) to surface model outputs to end users.
- Practical experience with Docker/containers and GPU compute for training/inference.
- Experience building and maintaining CI/CD pipelines for ML/AI workloads.
- Working knowledge of DevSecOps practices applied to ML pipelines.
- Experience providing operational support for production ML/AI systems, including monitoring and incident response.
- Experience implementing model governance and monitoring (drift detection, performance tracking, periodic retraining/tuning cycles).
- Demonstrated ability to design for human-in-the-loop / human-on-the-loop workflows for model oversight, retraining, and tuning.
- Demonstrated judgment in model/technique selection, including when to use AI/LLM approaches vs. traditional methods.
- Experience defining measurable testing/evaluation criteria for model performance and quality.
- Experience writing automated test cases, including using AI-assisted approaches to generate test coverage for model builds.
- Solid understanding of AI governance, legal, and security requirements, and experience embedding guardrails into model development.
- Familiarity with ML/AI frameworks (e.g., PyTorch, TensorFlow, Hugging Face, LangChain/LlamaIndex or similar agentic/RAG frameworks).
Preferred Qualifications
- Working knowledge of GCP and Azure ML/AI services.
- Experience with responsible AI toolkits (bias/fairness testing, model explainability).
- Certifications in AWS ML/AI or relevant cloud platforms.