Sr Data Scientist & AI Developer

Job Description

As a Sr Data Scientist and AI Developer here at Honeywell, you will play a crucial role in designing and implementing advanced data solutions for AI solutions that drive business insights, enhance decision‑making processes and empower AI solutions. Your expertise will help in critical AI development activities across all AI modalities (classic, Gen and agentic) and data types (structured and unstructured).

You will report directly to our AI Director and you’ll work out of our Phoenix, AZ location on a Hybrid work schedule.

Note:

New Hires will work onsite M‑F for the first 90 days. In this role, you will impact the organization by leveraging your technical skills to develop innovative software solutions that support strategic initiatives and improve operational efficiency.

• Design, develop, and deploy advanced machine learning models, LLM-based solutions, and agentic AI systems to solve complex business problems across diverse domains.
• Conduct exploratory data analysis, statistical assessments, and feature engineering on structured, semi‑structured, and unstructured datasets.
• Build and evaluate GenAI workflows including prompt engineering, fine‑tuning, RAG pipelines, embedding analysis, and context optimization.
• Develop and validate agentic AI behaviors, including reasoning chains, tool‑use strategies, action planning, memory utilization, and safety constraints.
• Partner with Data Engineers, AI Developers, Platform Engineers, and MLOps to bring models and agents into production using Databricks, Dataiku, MLflow, and AWS‑native deployment patterns.
• Develop robust evaluation frameworks for ML models, LLMs, and agentic systems—covering accuracy, robustness, hallucination resistance, safety, bias, reliability, and task success rate.
• Implement experiments, compare algorithms, perform ablation studies, and use statistical methods to quantify improvements for both classic ML and LLM‑based systems.
• Translate complex AI insights (predictions, feature impacts, agent decisions, retrieval context) into clear business recommendations and decision frameworks.
• Stay current with emerging trends in AI—including new model families, multi‑modal approaches, vector search innovations, and agentic frameworks—and assess applicability within the enterprise.
• Contribute to reusable AI assets such as feature stores, embedding stores, evaluation datasets, agent toolkits, and documentation playbooks.

• Bachelor’s degree from an accredited institution in a technical discipline such as science, technology, engineering, mathematics.
• 4–7 years of experience building, evaluating, and deploying machine learning models in production environments.
• Strong proficiency in Python and key ML/AI libraries (pandas, Num Py, scikit‑learn, PyTorch or Tensor Flow, Hugging Face Transformers).
• Applied experience developing LLM‑based solutions, including prompt engineering, retrieval‑augmented generation (RAG), embeddings, and evaluation.
• Experience working with Databricks (Spark, Delta Lake, Unity Catalog, MLflow) for data preparation, training, and experiment tracking.
• Experience with Dataiku for workflow orchestration, data pipelines, and model deployment/use in AI applications.
• Hands‑on experience with AWS data and AI services such as S3, Lambda, Step Functions, Glue, Bedrock, or Sage Maker.
• Strong statistical background with experience in hypothesis testing, regression, clustering, classification, and optimization techniques.
• Ability to communicate complex findings clearly to technical and non‑technical stakeholders.
• Proven ability to collaborate in cross‑functional agile teams, partnering with engineering, MLOps, and product owners.

• Bachelor’s or Master’s degree in Computer Science, Mathematics, Statistics, Engineering, or a related quantitative discipline.
• Experience with agentic AI systems, including:
  • Tool/function calling
  • Multi‑step reasoning evaluation
  • Memory and retrieval strategies
  • Human‑in‑the‑loop review patterns
  • Safety and guardrail testing
• Experience evaluating LLMs for accuracy, hallucination, chain‑of‑thought, content safety, and task reliability.
• Familiarity with vector databases (Databricks Vector…