Senior Full-Stack Developer / AI Engineer / Data Scientist (m/w/d)
Location: Berlin / Remote (DACH time zone)
Employment Type: Full-time, permanent
Start Date: Immediate
Language: Native-level English proficiency (C1-C2 level, German a bonus)
_____________________
ABOUT US
sandan AI is a Berlin-based AI-native startup serving enterprise customers across the DACH region. We are shaping what marketing looks like in an agentic economy — and building the platform on which this new generation of marketing workflows runs.
Our ambition is to build software the way software should be built in 2026: agentic, iterative, and with AI as an integral part of the engineering workflow.
HOW WE WORK
We use code harnesses such as Claude Code and Codex not as gimmicks, but as a core part of our development process. Our engineers work daily with agentic coding workflows — from architecture sketches and implementation to refactoring and testing.
People joining us either already work productively with these tools or bring the clear motivation to learn and master them quickly and deeply.
Tasks
Develop and maintain our full-stack applications within the modern JavaScript/TypeScript ecosystem
Design and implement scalable backend services using Node.js and Nest.js
Build high-performance frontend solutions with React.js and Next.js
Design and optimize relational databases (PostgreSQL, MySQL, MS SQL)
Containerize and orchestrate services using Docker and Kubernetes
Develop and integrate REST APIs
Build and operate data pipelines, feature stores, and ML workflows for our agentic systems
Develop, train, and evaluate machine learning models for marketing use cases (attribution, targeting, forecasting, anomaly detection)
Implement Retrieval-Augmented Generation (RAG) pipelines (embeddings, vector stores, re-ranking)
Set up and maintain LLM evaluation frameworks, prompt optimization workflows, and systematic quality benchmarks
Apply code harnesses (especially Claude Code and Codex) throughout the daily development process — from planning and implementation to code review
Collaborate closely with the founding team on the product architecture of agentic systems
Requirements
Core Technical Competencies (Engineering)
Several years of professional experience in full-stack development
Strong knowledge of TypeScript and JavaScript
Confident working with React.js, Next.js, Node.js, and Nest.js
Experience with relational databases (PostgreSQL, MySQL, MS SQL)
Hands-on experience with container technologies (Docker, Kubernetes)
Familiarity with REST API design and Git workflows
Data Science & Machine Learning
Solid Python skills, including common data/ML stacks (pandas, NumPy, scikit-learn, PyTorch, or TensorFlow)
Experience across the end-to-end lifecycle of ML models: data preparation, feature engineering, training, validation, deployment, and monitoring
Strong understanding of statistical fundamentals (hypothesis testing, regression, classification, time series)
Experience with modern LLM workflows: prompt engineering, RAG, embeddings, vector databases (e.g. Pinecone, Weaviate, pgvector)
Hands-on experience in model evaluation: metrics, benchmarks, A/B testing, and systematic evaluation frameworks
Experience with notebook-based exploration environments (Jupyter, Colab) and transitioning them into production-ready pipelines
Familiarity with MLOps tools (e.g. MLflow, Weights & Biases, DVC) is a plus
Code Harnesses
Experience with Claude Code and/or Codex is highly appreciated — otherwise, a strong willingness to actively learn and adopt these tools
Personal Qualifications
Excellent German (C2) and English skills (minimum C1)
Genuine interest in generative AI and agentic systems
Builder mentality: you deliver production-grade software, not just prototypes
Analytical mindset: you make data-driven decisions and critically evaluate model behavior
Nice to have
A defining role in an AI-native startup with a strong enterprise focus
Competitive salary and employee stock ownership plan (ESOP)
Fully funded access to all relevant code harnesses, LLM APIs, and developer tools (Claude Max, Codex, GitHub Copilot, etc.)
Compute budget for experiments, model training, and ML workloads
Modern hardware setup (MacBook Pro of your choice)
Flexible working hours and remote work options
Office in Berlin and location in Dubai (DIFC AI Campus)
A steep learning curve in Agentic AI, LLM orchestration, and AI-native engineering
Benefits
A defining role in an AI-native startup with a strong enterprise focus
Competitive salary and employee stock ownership plan (ESOP)
Fully funded access to all relevant code harnesses, LLM APIs, and developer tools (Claude Max, Codex, GitHub Copilot, etc.)
Compute budget for experiments, model training, and ML workloads
Modern hardware setup (MacBook Pro of your choice)
Flexible working hours and remote work options
Office in Berlin and location in Dubai (DIFC AI Campus)
A steep learning curve in Agentic AI, LLM orchestration, and AI-native engineering
We look forward to receiving your application, including your CV and — if available — a GitHub or portfolio link.
Please also let us know your earliest possible start date.
Bonus: Tell us in two or three sentences about the most recent ML project you worked on — and how you currently use Claude Code, Codex, or comparable tools in your workflow.
Apply To This Job
Employment Type: Full-time, permanent
Start Date: Immediate
Language: Native-level English proficiency (C1-C2 level, German a bonus)
_____________________
ABOUT US
sandan AI is a Berlin-based AI-native startup serving enterprise customers across the DACH region. We are shaping what marketing looks like in an agentic economy — and building the platform on which this new generation of marketing workflows runs.
Our ambition is to build software the way software should be built in 2026: agentic, iterative, and with AI as an integral part of the engineering workflow.
HOW WE WORK
We use code harnesses such as Claude Code and Codex not as gimmicks, but as a core part of our development process. Our engineers work daily with agentic coding workflows — from architecture sketches and implementation to refactoring and testing.
People joining us either already work productively with these tools or bring the clear motivation to learn and master them quickly and deeply.
Tasks
Develop and maintain our full-stack applications within the modern JavaScript/TypeScript ecosystem
Design and implement scalable backend services using Node.js and Nest.js
Build high-performance frontend solutions with React.js and Next.js
Design and optimize relational databases (PostgreSQL, MySQL, MS SQL)
Containerize and orchestrate services using Docker and Kubernetes
Develop and integrate REST APIs
Build and operate data pipelines, feature stores, and ML workflows for our agentic systems
Develop, train, and evaluate machine learning models for marketing use cases (attribution, targeting, forecasting, anomaly detection)
Implement Retrieval-Augmented Generation (RAG) pipelines (embeddings, vector stores, re-ranking)
Set up and maintain LLM evaluation frameworks, prompt optimization workflows, and systematic quality benchmarks
Apply code harnesses (especially Claude Code and Codex) throughout the daily development process — from planning and implementation to code review
Collaborate closely with the founding team on the product architecture of agentic systems
Requirements
Core Technical Competencies (Engineering)
Several years of professional experience in full-stack development
Strong knowledge of TypeScript and JavaScript
Confident working with React.js, Next.js, Node.js, and Nest.js
Experience with relational databases (PostgreSQL, MySQL, MS SQL)
Hands-on experience with container technologies (Docker, Kubernetes)
Familiarity with REST API design and Git workflows
Data Science & Machine Learning
Solid Python skills, including common data/ML stacks (pandas, NumPy, scikit-learn, PyTorch, or TensorFlow)
Experience across the end-to-end lifecycle of ML models: data preparation, feature engineering, training, validation, deployment, and monitoring
Strong understanding of statistical fundamentals (hypothesis testing, regression, classification, time series)
Experience with modern LLM workflows: prompt engineering, RAG, embeddings, vector databases (e.g. Pinecone, Weaviate, pgvector)
Hands-on experience in model evaluation: metrics, benchmarks, A/B testing, and systematic evaluation frameworks
Experience with notebook-based exploration environments (Jupyter, Colab) and transitioning them into production-ready pipelines
Familiarity with MLOps tools (e.g. MLflow, Weights & Biases, DVC) is a plus
Code Harnesses
Experience with Claude Code and/or Codex is highly appreciated — otherwise, a strong willingness to actively learn and adopt these tools
Personal Qualifications
Excellent German (C2) and English skills (minimum C1)
Genuine interest in generative AI and agentic systems
Builder mentality: you deliver production-grade software, not just prototypes
Analytical mindset: you make data-driven decisions and critically evaluate model behavior
Nice to have
A defining role in an AI-native startup with a strong enterprise focus
Competitive salary and employee stock ownership plan (ESOP)
Fully funded access to all relevant code harnesses, LLM APIs, and developer tools (Claude Max, Codex, GitHub Copilot, etc.)
Compute budget for experiments, model training, and ML workloads
Modern hardware setup (MacBook Pro of your choice)
Flexible working hours and remote work options
Office in Berlin and location in Dubai (DIFC AI Campus)
A steep learning curve in Agentic AI, LLM orchestration, and AI-native engineering
Benefits
A defining role in an AI-native startup with a strong enterprise focus
Competitive salary and employee stock ownership plan (ESOP)
Fully funded access to all relevant code harnesses, LLM APIs, and developer tools (Claude Max, Codex, GitHub Copilot, etc.)
Compute budget for experiments, model training, and ML workloads
Modern hardware setup (MacBook Pro of your choice)
Flexible working hours and remote work options
Office in Berlin and location in Dubai (DIFC AI Campus)
A steep learning curve in Agentic AI, LLM orchestration, and AI-native engineering
We look forward to receiving your application, including your CV and — if available — a GitHub or portfolio link.
Please also let us know your earliest possible start date.
Bonus: Tell us in two or three sentences about the most recent ML project you worked on — and how you currently use Claude Code, Codex, or comparable tools in your workflow.
Apply To This Job