What you will do
We are hiring a Senior Data Scientist to play a key role in building our Agentic AI Platform a system of autonomous, tool-using AI agents that plan, reason, and execute complex business workflows end-to-end. The ideal candidate combines strong ML fundamentals with hands-on experience in LLM-based application development, agent orchestration frameworks, and Microsoft Azure cloud services. You will architect and ship production-grade agentic solutions, mentor junior team members, and set technical direction for GenAI initiatives across the organization.
How you will do it
Design, build, and productionize multi-agent systems, including planning, tool calling / function calling, memory, and orchestration, using frameworks such as LangGraph, AutoGen, CrewAI, or Semantic Kernel.
Develop RAG pipelines end-to-end: document ingestion, chunking strategies, embeddings, vector search (Azure AI Search / FAISS / pgvector), re-ranking, and grounding for agent knowledge.
Integrate agents with enterprise systems and APIs via tool/function calling and Model Context Protocol (MCP) or similar connector patterns.
Build and maintain LLM evaluation frameworks for agentic workflows, task-completion metrics, hallucination detection, trajectory analysis, LLM-as-judge pipelines, and A/B testing.
Implement guardrails, safety, and governance for agents: prompt-injection defense, content filtering, role-based tool permissions, human-in-the-loop checkpoints, and audit logging.
Fine-tune and optimize LLMs where needed (LoRA/PEFT, prompt optimization, model routing, latency/cost trade-offs) on Azure OpenAI / Azure AI Foundry.
Design, build, and evaluate classical ML models (classification, regression, forecasting, NLP) where they complement agentic workflows.
Own LLMOps/MLOps for the platform: experiment tracking, prompt versioning, CI/CD, observability and tracing (LangSmith, Azure Monitor, OpenTelemetry), drift monitoring, and retraining strategies.
Collaborate with data engineers on data quality, availability, and governance across Azure Data Lake, Databricks, and Synapse Analytics.
Translate ambiguous business problems into agentic AI solutions; present architecture decisions and results to senior stakeholders.
Mentor junior data scientists, lead code/design reviews, and champion engineering best practices.
Required
professional experience in data science / ML, with building LLM or GenAI applications in production.
Hands-on experience with agentic frameworks LangChain or LangGraph.
Strong understanding of LLM application patterns including prompt engineering
Expert-level Python (pandas, NumPy, scikit-learn, async programming, API development with FastAPI).
Experience with vector databases and embeddings (Azure AI Search, Pinecone, Weaviate, Qdrant, FAISS, or pgvector).
Strong SQL skills and experience working with large-scale data.
Preferred
Bachelor's or Master's degree in Computer Science, Data Science, Engineering, Mathematics, or a related field.
Hands-on experience with Azure services: Azure OpenAI / AI Foundry, Azure ML, Azure AI Search, Azure Databricks, Azure Data Factory, or Synapse.
Solid grounding in classical ML: supervised/unsupervised learning, model evaluation, and hyperparameter tuning; experience with TensorFlow or PyTorch.
Experience with Model Context Protocol (MCP), OpenAI Assistants/Agents SDK, or Anthropic tool-use APIs.
Strong understanding of function/tool calling, structured outputs, RAG, agent memory, and multi-agent orchestration.
Proven MLOps/LLMOps experience: MLflow, prompt/model versioning, CI/CD (Azure DevOps or GitHub Actions), and production monitoring.
Ability to evaluate and mitigate LLM-specific risks: hallucination, prompt injection, data leakage, and cost/latency constraints.
Experience with AutoGen, CrewAI, Semantic Kernel.
Microsoft certifications: AI-102 (Azure AI Engineer), DP-100 (Azure Data Scientist Associate).
Experience fine-tuning open-source LLMs (Llama, Mistral, Phi) using LoRA/QLoRA and serving via vLLM or Azure ML endpoints.
Familiarity with observability/tracing for agents: LangSmith, Langfuse, Arize Phoenix, or OpenTelemetry.
Knowledge of containerization and deployment: Docker, Kubernetes (AKS), Azure Container Apps.
Big data experience with Apache Spark (PySpark) via Azure Databricks.
Experience with knowledge graphs, graph RAG, or semantic layers for agent grounding.
Contributions to open-source GenAI/agentic projects or published technical content.
Our salary and benefits
The initial basic salary for this position will be in the range of 69,966 CZK - 109 333 CZK per month, plus a 13th month salary applied after your first 6 months in the job. The final package will reflect your experience, skills, and qualifications relevant to the role. We are committed to fair, equitable, and gender-neutral pay practices.
At Johnson Controls, your work comes with benefits that support both your career and your wellbeing.
Meal vouchers fully covered by the company.
Flexible benefits budget with access to 3,500+ options, including a MultiSport card, medical and wellness services.
Extra savings through retail and lifestyle discounts (Benefit+).
Allowance for the private kindergarten or nursery to support your family.
Flexible working hours and home office days.
Allowance for language courses, professional development support, and wellbeing support.
On site benefits: massages, distribution of fruits, yoga, psychologist, health month, various events with supporting groups (volunteering activities, branding activities).
Seniority benefits - extra monthly financial allowance, medical care and support, wellbeing day.
Relocation support with housing allowance if moving to Slovakia
About Us
Johnson Controls, a global leader in thermal management, mission-critical building systems, energy efficiency, and decarbonization, helps customers use energy more productively, reduce carbon emissions, and operate with the precision and resilience required in rapidly expanding industries such as data centers, healthcare, pharmaceuticals, advanced manufacturing, and higher education.
For more than 140 years, Johnson Controls has delivered performance where it really matters. Backed by advanced technology, lifecycle services and an industry-leading field organization, we elevate customer performance, turn goals into real-world results and help move society forward.
We are committed to diversity and inclusion and believe that different perspectives make us stronger. By encouraging open dialogue and valuing individuality, we strive to be one of the most desirable places to work.
#LI-Hybrid
#LI-JT1
Skills Required
- Professional experience in data science or machine learning
- Production experience building LLM or GenAI applications
- Hands-on experience with LangChain or LangGraph
- Strong understanding of prompt engineering and LLM application patterns
- Expert-level Python, including pandas, NumPy, scikit-learn, asynchronous programming, and FastAPI
- Experience with vector databases and embeddings
- Strong SQL skills and experience with large-scale data
- Bachelor's or Master's degree in Computer Science, Data Science, Engineering, Mathematics, or a related field
- Experience with Azure OpenAI, Azure AI Foundry, Azure Machine Learning, Azure AI Search, Azure Databricks, Azure Data Factory, or Synapse
- Knowledge of supervised and unsupervised learning, model evaluation, and hyperparameter tuning
- Experience with TensorFlow or PyTorch
- Experience with Model Context Protocol, OpenAI Assistants or Agents SDK, or Anthropic tool-use APIs
- Understanding of function calling, structured outputs, RAG, agent memory, and multi-agent orchestration
- MLOps or LLMOps experience, including MLflow, versioning, CI/CD, and production monitoring
- Ability to evaluate and mitigate hallucinations, prompt injection, data leakage, cost, and latency risks
- Experience with AutoGen, CrewAI, or Semantic Kernel
- Microsoft AI-102 or DP-100 certification
- Experience fine-tuning open-source LLMs with LoRA or QLoRA and serving them via vLLM or Azure ML endpoints
- Familiarity with agent observability and tracing tools
- Knowledge of Docker, Kubernetes, AKS, or Azure Container Apps
- Big data experience with Apache Spark or PySpark via Azure Databricks
- Experience with knowledge graphs, graph RAG, or semantic layers
- Contributions to open-source GenAI or agentic projects, or published technical content
Johnson Controls Compensation & Benefits Highlights
The following summarizes recurring compensation and benefits themes identified from responses generated by popular LLMs to common candidate questions about Johnson Controls and has not been reviewed or approved by Johnson Controls.
-
Retirement Support — Retirement support is positioned as a meaningful part of the package through employer 401(k) matching, repeatedly framed as a strong pillar of the overall rewards mix. The matching contribution is described with specific match levels in multiple places, reinforcing perceived value for long-term saving.
-
Leave & Time Off Breadth — Time off is presented as comparatively robust, with multiple paid holiday categories, vacation time, and sick time described as generous or “amazing” in places. Paid time off breadth appears to be a consistent contributor to total rewards attractiveness beyond base pay.
-
Flexible Benefits — Benefits are described as broad and customizable, spanning standard medical/dental/vision plus optional add-ons like pet insurance, identity protection, and legal support. Tuition reimbursement is repeatedly highlighted as a high-value option supporting professional development.
Johnson Controls Insights
What We Do
At Johnson Controls, we transform the environments where people live, work, learn and play. From optimizing building performance to improving safety and enhancing comfort, we drive the outcomes that matter most. Dedicated to protecting the environment, we deliver our promise in industries such as healthcare, education, data centers and manufacturing. With a global team of 100,000 experts in more than 150 countries and over 130 years of innovation, we are the power behind our customers’ mission. Our leading portfolio of building technology and solutions includes some of the most trusted names in the industry, such as Tyco®, York®, Metasys®, Ruskin®, Titus®, Frick®, Penn®, Sabroe®, Simplex®, Ansul® and Grinnell®.
