דרושים » תוכנה » Senior ML Data Engineer

We are looking for someone who brings ML and computer-vision depth to the team - someone who can help shape the intelligence layer that decides what data is worth training on.
What will your job look like:
Work collaboratively with shared ownership. Your focus area will be the curation and ML side of our data pipeline, but you will contribute across the full stack alongside the rest of the team.
Build and improve the curation pipeline - from vision-model embeddings and scene detection, through VLM-based scene analysis, to scoring, deduplication, and sampling that produces a balanced and diverse dataset.
Run and optimize GPU inference at scale (embedding extraction, VLM inference) across thousands of driving sessions using workflow orchestration.
Develop scoring and sampling strategies that ensure rare but important scenarios (night driving, adverse weather, hazardous situations) are well-represented in the final dataset.
Work with algorithm teams to understand what data gaps hurt model performance and translate those into curation criteria.
Build validation and diagnostics that measure dataset quality - not just pipeline health, but whether the data is actually good for training.
Contribute to the core dataset SDK, converter, and 3D-geometry tooling (camera projection, calibration, coordinate transforms).

We're looking for a Data Engineer with 4+ year experience to join our Data Engineering team and help us build and scale our production-grade data platform. You'll work on high-performance systems built on self-hosted ClickHouse, optimize complex data pipelines, and collaborate closely with Product, Analytics, and Infrastructure teams to deliver reliable, fast, and scalable data solutions.

This is a hands-on technical role where you'll have a significant impact on how we ingest, model, store, and serve data that powers our analytics and AI-driven products.
Youll play a key role in shaping the direction of our data platform and have meaningful ownership over critical components of our architecture.

What You'll Do:
Data Modeling & Architecture
Design and evolve data models that reflect business logic and support analytical use cases
Collaborate with the BI and Analytics teams to understand data requirements and translate them into efficient schemas
Performance Optimization
Optimize ClickHouse schemas, partitioning strategies, indexing, and compression
Profile and tune slow queries to improve performance and reduce costs
Implement systems that ensure data quality, consistency, and operational efficiency (e.g., deduplication, validation, anomaly detection)
Monitor pipeline health, data freshness, and query performance with appropriate alerting mechanisms
SQL Compiler Development
Develop and maintain the SQL Compiler layer that translates high-level queries into optimized ClickHouse execution plansImplement query optimization and rewriting strategies to improve performanceDebug and resolve compiler issues to ensure accurate and efficient query translation

Data Pipeline Development & Collaboration
Review and advise the Integration team on pipeline architecture, performance, and best practices.
Provide guidance on data modeling, schema design, and optimization for new data sources.
Troubleshoot and maintain existing pipelines when issues arise or optimization is needed
Ensure data freshness, reliability, and quality across all ingestion pipelines.
Collaboration & Support
Work closely with the Integration team to ensure smooth data ingestion from new sources.
Partner with Infrastructure to support high availability and disaster recovery
Support other teams across the company in accessing and using data effectively.

As an ML Software Engineer with a focus on low-level and CUDA-based optimizations, you will play a key role in shaping the design, performance, and scalability of machine learning inference systems. Youll work on deeply technical challenges at the intersection of GPU acceleration, systems architecture, and ML deployment.
Your expertise in CUDA, C/C++, and performance tuning will be crucial in enhancing runtime efficiency across heterogeneous computing environments. Youll collaborate with designers, researchers, and backend engineers to build production-grade ML pipelines that are optimized for latency, throughput, and memory use, contributing directly to the infrastructure powering next-generation AI products.This role is ideal for an engineer with strong systems-level thinking, deep familiarity with GPU internals, and a passion for pushing the boundaries of performance and efficiency in machine learning infrastructure.

Responsibilities
Design and implement highly optimized GPU-accelerated ML inference systems using CUDA and low-level parallelism techniques
Optimize memory, compute, and data flow to meet real-time or high-throughput constraints
Improve the performance, reliability, and observability of our inference backend across diverse compute targets (CPU/GPU)
Collaborate with cross-functional teams (including researchers, developers, and designers) to deliver efficient and scalable inference solutions
Contribute to ComfyUI and internal infrastructure to improve usability and performance of model execution flows
Investigate performance bottlenecks at all levels of the stack-from Python to kernel-level execution
Navigate and enhance a large, complex, production-grade codebase
Drive innovation in low-level system design to support future ML workloads

The Model Evaluation team is the central nervous system of the LTX Foundation Model group. We don't just measure performance; we define what "good" looks like across a vast array of use cases. While we power the next generation of creative tools, LTX is also a foundational engine for simulation pipelines, game engines, synthetic data generation, architectural rendering, and digital avatars. We act as the critical bridge between raw research and industrial-grade reliability, building the benchmarks that ensure our models are world-class for both artists and engineers.

The Role
As a Research Scientist in Model Evaluation, you are the ultimate authority on model quality and utility. You will design the automated judges, reward models, evaluation datasets, and benchmarking ecosystems that determine the future of LTX. Your mission is to provide the "ground truth" for our pre-training and post-training teams. You will blend the rigor of a researcher with the intuition of a product-thinker, developing metrics that capture both the aesthetic soul of a video and the functional precision required for high-stakes professional use.

Key Responsibilities
Steer Training & Research: Systematically evaluate model checkpoints to provide actionable insights that guide training experiments and architectural decisions.
Design Benchmark Ecosystems: Develop and run rigorous benchmarks for release candidates against competitive models, ensuring LTX-2 remains world-class.
Build Next-Gen Metrics: Develop robust automatic metrics and Reward Models (e.g., for RL, ITS, auto-research agents) that quantify complex attributes like temporal coherence, physical correctness, spatial accuracy, and foley synchronization.
Diagnose & Analyze: Perform deep root-cause analysis on model failures, providing the diagnostic clarity needed for researchers to implement targeted fixes.
Scale Evaluation: Collaborate with platform engineers to deploy evaluation frameworks across large-scale GPU clusters.

Our CTO Group is looking for an outstanding Physical-AI Applied Researcher to join our team.
The CTO Group is a small, elite research unit shaping the next generation of algorithmic foundations behind our autonomous driving systems. The group operates at the core of the decision-making and planning stack, addressing some of the most challenging problems in real-world autonomy.
We are seeking a researcher who thrives at the intersection of machine learning, decision-making, and algorithmic rigor - someone who is excited about advancing learning-based approaches for safety-critical, large-scale physical systems.
In this role, you will develop novel approaches for planning and decision-making in interactive, multi-agent driving environments. You will combine deep & reinforcement learning with classical algorithmic structure and formal reasoning. The problems are open-ended, scientifically challenging, and deployed at unprecedented scale.
This is a rare opportunity to conduct high-impact applied research, taking ideas from theory and papers into real-world autonomous systems at scale. If youre excited about pushing the boundaries of learning-based decision-making, wed love you to join us and help shape the future of Physical AI.
What will your job look like:
Design and develop novel learning-based algorithms for decision-making and planning in complex physical environments.
Advance model architectures for long-horizon reasoning, multi-agent interaction, and uncertainty-aware prediction.
Integrate deep learning components into structured planning pipelines with clear formal objectives and safety constraints.
Formulate problems mathematically and derive principled learning objectives grounded in real-world system requirements.
Lead research directions from conception to full-scale production.
Develop using Python (PyTorch or similar frameworks) as well as C++/GPU/Cuda.

This role is designed for individuals who are not only technically proficient but also deeply passionate about pushing the boundaries of AI and machine learning through innovative engineering and collaborative research.

Key Responsibilities
Profile and optimize the training process to ensure efficiency and effectiveness, including optimizing multimodal data pipelines and data storage methods.
Develop high-performance TPU/GPU/CPU kernels and integrate advanced techniques into our training framework to maximize hardware efficiency.
Utilize knowledge of hardware features to make aggressive optimizations and advise on hardware/software co-designs.
Collaboratively develop model architectures with researchers that facilitate efficient training and inference.
Design, maintain, and evolve a high-quality, shared codebase that emphasizes correctness, readability, extensibility, testing, and long-term maintainability, while balancing performance requirements.

We are seeking a Backend & Data Engineer to join our Innovation Team within our Mapping Division. This role is best suited for engineers with strong system-level thinking, a can-do approach, and a hands-on mindset, with the ability to design, build, and optimize complex systems operating at scale.
What your job will look like:
Develop and maintain backend and data-processing components in large-scale systems
Design, implement, and optimize data pipelines and distributed processing flows
Work with large-scale storage systems (e.g., S3) and high-volume data access patterns
Optimize systems and code across multiple layers - from architecture to implementation
Identify performance bottlenecks, debug complex issues, and drive root-cause solutions
Work across teams and domains, reading, improving, and refactoring existing code
Take part in technical design and decision-making, balancing performance, scalability, and maintainability.

This role focuses on pioneering model architecture and pre-training algorithms, shaping the next generation of our foundational generative AI models.

What you will be doing
Pre-train and fine-tune video, audio, and image generative models to pursue state-of-the-art results.
Publish papers and open source models to benefit the research community and advance the field.
Design and implement machine learning models for text-to-audio and text-to-video generation.
Collaborate with data engineers to curate and preprocess text and video data.
Optimize models for high performance, ensuring efficient training and inference.
Build new controls and capabilities into generative text-to-audio and text-to-video models.
Stay updated with the latest developments in Generative AI, particularly in the fields of image, video, and audio.
Work closely with product teams to integrate AI models into applications and services.
Conduct experiments and prototype new concepts to advance the capabilities of our AI tools.

As the ML Platform Tech Lead, youll be responsible for the development of AI infrastructure, taking an active role in mentoring backend engineers directly.

This is a hands-on tech- lead position that blends system design and technical strategy. You will be expected to work closely with our ML research teams, product, and engineering stakeholders to take cutting-edge AI models into production, powering features across our creative apps used by millions worldwide.

You will own the full development lifecycle, from system architecture through deployment and operation, while also being accountable for team delivery, collaboration, and growth.

As a Large Scale Video Understanding Research Scientist, you will play a key role in improving video generation quality and efficiency by improving video and audio understanding pipelines used for both training data construction and model evaluation.. This role demands hands-on work with large-scale Video Language Models (VLLMs), including fine-tuning, post-training, and control, alongside implementing classic computer vision and signal processing algorithms and applying strong research skills. Your expertise in post-training and controlling large scale foundational models, understanding statistics, implementing complex systems and eliminating bugs will be crucial, as our video training sets consist of petabytes of data processed across hundreds to thousands of virtual machines.

What you will be doing
Fine-tune and control VLLMs for video and audio understanding.
Design algorithms for balancing, filtering, and curating training and evaluation datasets, informed by model behavior and failure modes.
Implement classic and modern algorithms for processing, clustering, evaluation and filtering of large scale datasets.
Work within high-performance, scalable distributed systems capable of handling petabytes of data, with attention to throughput, correctness, and reproducibility..
Collaborate with other researchers and product stakeholders to iteratively improve training sets and evaluation protocols through tight feedback loops driven by model performance.