Data Scientist - Agentic AI Developer

Data Scientist - Agentic AI Developer

Req number:

R8813

Employment type:

Full time

Worksite flexibility:

Remote

##

Who we are

CAI is a global services firm with over 9,000 associates worldwide and a yearly revenue of $1.3 billion+. We have over 40 years of excellence in uniting talent and technology to power the possible for our clients, colleagues, and communities. As a privately held company, we have the freedom and focus to do what is right—whatever it takes. Our tailor-made solutions create lasting results across the public and commercial sectors, and we are trailblazers in bringing neurodiversity to the enterprise.

Job Summary

We are looking for a motivated Data Scientist - Agentic AI Developer ready to take us to the next level! If you have advanced Python, Databricks, and AWS Bedrock experience and are looking for your next career move, apply now

Job Description

We are looking for a Data Scientist - Agentic AI Developer to develop, deploy, and monitor production-ready machine learning, Generative AI, LLM, RAG, and Agentic AI solutions. This position will be full-time and remote.

What You’ll Do

• Partner with business and technical stakeholders to define problems, success metrics, KPIs, constraints, and solution approaches
• Perform data sourcing, exploratory data analysis, statistical analysis, feature engineering, data-quality checks, and hypothesis testing
• Build and evaluate machine learning, deep learning, time-series, anomaly-detection, recommendation, and optimization models
• Develop production-ready LLM, RAG, and Agentic AI applications using LangGraph, LangChain, Amazon Bedrock Agents, or similar frameworks
• Design multi-step and multi-agent workflows involving tool calling, state and memory management, human-in-the-loop approvals, retries, fallbacks, tracing, and MCP integrations
• Build secure and scalable AI solutions using Databricks Apps, MLflow, Model Serving, Vector Search, Workflows, Delta Lake, and Unity Catalog
• Develop applications on Amazon Bedrock using foundation models, Knowledge Bases, Agents, Guardrails, prompt management, and model evaluation
• Design advanced RAG pipelines involving hybrid retrieval, metadata filtering, re-ranking, query rewriting, embeddings, vector databases, and retrieval evaluation
• Build NLP solutions for text classification, named entity recognition, information extraction, semantic search, and document summarization
• Develop computer vision solutions for classification, object detection, segmentation, OCR, and visual defect inspection
• Fine-tune LLMs, embedding models, and vision models using LoRA, QLoRA, PEFT, and Hugging Face frameworks
• Develop and deploy AI services using FastAPI, REST APIs, Docker, Lambda, API Gateway, ECS, EKS, and CI/CD pipelines
• Monitor model performance, drift, quality, latency, reliability, and cost
• Conduct error analysis, model evaluation, and continuous improvement based on technical and business metrics
• Communicate technical findings, trade-offs, risks, and recommendations to business and engineering stakeholders

What You’ll Need

Required:

• 6 to 8 years of Data Science experience, including at least 3 years of hands-on Generative AI experience
• At least 5 years of advanced Python development experience
• Strong expertise in Python, Pandas, NumPy, PySpark, SQL, and object-oriented programming
• Strong foundations in statistics, probability, machine learning, deep learning, hypothesis testing, and model evaluation
• Hands-on experience with Scikit-learn, XGBoost, LightGBM, and PyTorch or TensorFlow
• Experience building models and AI applications from first principles rather than relying only on AutoML or pre-built components
• Hands-on experience building LLM applications using Claude through Amazon Bedrock or other foundation models
• Strong understanding of prompt engineering, structured outputs, function and tool calling, context management, caching, and LLM evaluation
• Proven experience building at least one end-to-end multi-step or multi-agent system
• Experience with agent tool design, state and memory management, human-in-the-loop workflows, retries, fallbacks, evaluation, tracing, and MCP
• Experience with RAG, embeddings, vector databases, hybrid search, metadata filtering, re-ranking, and query rewriting
• Knowledge of retrieval evaluation metrics including recall@k, MRR, and nDCG
• Hands-on experience with Databricks Apps, MLflow, Model…