ADAS Analytics ML Scientist: Vision & LLM Pipelines

Mercedes-Benz Research & Development North America, Inc.; San Jose (CA); On-site; USD 118,000 - 148,000; Full time

Benefits offered by this job

Medical, dental, vision insurance

401(k) with employer match

Up to 15 company-paid holidays

Paid time off and parental leave

Tuition assistance

Job summary

Mercedes-Benz Research & Development North America, Inc. is seeking a data scientist to scale ADAS capabilities using vision-language models, embeddings, and LLM-based classification.

You will build end-to-end pipelines, deploy models in production, and collaborate with engineers and management on fleet-scale analyses. The role emphasizes translating ambiguous questions into structured analytics, operating on large datasets, and delivering actionable insights with robust validation and

Qualifications

• Bachelor's or Master's in Data Science, ML, Statistics, Computer Science or related field
• 3+ years of professional experience
• 2-5 years in data science or applied machine learning
• Depth in foundation models, VLMs, or embeddings or traditional ML in production
• Strong Python with NumPy, pandas, scikit-learn and strong SQL
• Experience with ML model development including feature engineering and evaluation
• Strong statistics knowledge: hypothesis testing, regression, experiments
• Ability to communicate findings to engineers and management
• Ability to turn ambiguous questions into structured analytical approaches

Responsibilities

• Apply vision-language models to SSR video and combine with telemetry to create multimodal event representations
• Build LLM classification and reasoning pipelines; triage events by severity and provide summaries
• Design embedding pipelines and semantic search for similar-event retrieval and clustering
• Develop model-based ranking and active-learning loops; detect fleet-level anomalies
• Quantify impact of firmware updates on KPIs; segment driver cohorts and run AB/quasi experiments
• Analyze campaign effectiveness and develop fleet-quality scoring
• Deploy models into PySpark/Delta Lake pipelines and FastAPI analytics API; operate on Azure data platform

Skills

Foundation models (LLMs/VLMs/embeds)

Python (NumPy, pandas, scikit-learn)

SQL

Statistics & experimental design

Communication to engineers/management

Structured analytics approach

Machine learning model development

Change detection & anomaly analysis

Time-series/forecasting basics

LLM/VLM application experience

Education

Bachelor's or Master's in Data Science, ML, Statistics, CS

PhD beneficial but not required

Tools

PySpark

Delta Lake

FastAPI

MLflow

Weights & Biases

Azure

AWS/GCP

Job description

Mercedes-Benz Research & Development North America, Inc. is seeking a data scientist to scale ADAS capabilities using vision-language models, embeddings, and LLM-based classification.

You will build end-to-end pipelines, deploy models in production, and collaborate with engineers and management on fleet-scale analyses. The role emphasizes translating ambiguous questions into structured analytics, operating on large datasets, and delivering actionable insights with robust validation and