Data Flywheel Engineer - Tech lead

Neolix; Mountain View (CA); On-site; USD 300,000 - 500,000; Full time

Benefits offered by this job

Premium health

Dental insurance

Vision insurance

401(k)

Job summary

Neolix in Silicon Valley seeks a Data Flywheel Engineer to convert petabytes of real driving data into high-quality model fuel, creating the company’s deepest moat. This is core, not support.

You will own end-to-end data closed loops, build automated pipelines for multimodal time-series data, and drive down annotation costs while collaborating across training infra, controls, and safety teams to improve on-road metrics across the fleet.

Qualifications

• Experience with data mining, active learning, auto-labeling, and data-closed-loop automation.
• Strong ability to build large-scale data pipelines.
• Experience with multimodal time-series data (video, LiDAR, radar, trajectory).
• Ability to reason about annotation cost structures and automation paths.

Responsibilities

• Own the end-to-end data closed loop from mining to training feedback.
• Build automated pipelines for large-scale multimodal data.
• Increase auto-labeling ratio and manage cost structures.
• Lead hard-example mining and corner-case discovery.
• Collaborate with training infra, controls/E2E, and safety teams to impact on-road metrics.

Skills

Data intuition

Active learning

Auto-labeling

Data pipelines

Multimodal data

Cross-functional collaboration

Job description

Strategically the single most important role on this team.

About Neolix

We are building a Silicon Valley frontier autonomous-driving team for Neolix, the global leader in L4 autonomous driving. Before you read further, here is the scale we operate at:

• Massive Fleet: 30,000+ proprietary autonomous logistics vehicles (RoboVans) deployed across 300+ cities in 15+ countries.
• Proven Mileage: ~200 million autonomous kilometers logged on public roads.
• #1 Globally: In June 2026, Neolix ranked #1 globally in last-mile autonomous delivery on the Road to Autonomy Index (scoring 74.7, ahead of Starship, Serve Robotics, and Coco). This index is built by Autnmy AI and S&P Dow Jones Indices based on verified operational records and regulatory disclosures—not self‑reported marketing.

About the Role

Our fleet already runs at a massive scale on public roads. The data flywheel is the engine that turns that real-world scale into a compounding model advantage.

As our Data Flywheel Engineer, you will own the loop that converts petabytes of real operational driving data into high-quality model fuel—and into the company's deepest moat.

Note: This is not a support function. It is the core competitive advantage of an autonomous‑driving company operating at fleet scale.

What You'll Do

• Own the End-to-End Data Closed Loop: Drive mining, active learning, auto‑labeling, curation, and feedback into training—turning fleet data directly into model improvements.
• Build Automated Pipelines: Architect pipelines for large‑scale multimodal time‑series data (video / LiDAR / radar / trajectory), handling cleaning, alignment, storage, retrieval, and versioning.
• Drive Down Annotation Costs: Radically increase the auto‑labeling ratio. You must be able to reason clearly about the human‑vs‑model‑vs‑rule cost structure (targeting benchmarks like Momenta's ~99% automation).
• Surface the Long‑Tail: Lead hard‑example mining and corner‑case discovery, closing the loop so that rare real‑world events become vital training signals.
• Collaborate for On‑Road Impact: Partner directly with training‑infra, controls/E2E, and safety teams to ensure the flywheel lifts critical on‑road metrics (MPI, interventions, safety) across the production fleet.

What We Look For

• Exceptional Data Intuition (The Soul of this Role): You have a real feel for data quality, distribution, and the long‑tail. You know exactly which slice of data actually pushes the model's ceiling.
• Hands‑On Expertise: Deep experience with data mining, active learning, auto‑labeling, and data‑closed‑loop automation.
• Large‑Scale Data Engineering: Fluency in building large‑scale data pipelines using Spark, Ray, and stream‑batch architectures, along with data versioning and lineage.
• Multimodal Mastery: Proven ability to work with multimodal time‑series data (video / LiDAR / radar / trajectory).
• Strategic Vision: A clear understanding of annotation cost structures and the technical path to full automation.
• Previously worked at an industry‑leading autonomous driving company (e.g., Waymo, Tesla, Cruise, or Nuro)

Ideal Backgrounds

You likely come from a data‑flywheel benchmark team (Momenta, Tesla Autopilot, Waymo, XPeng), an elite data‑engine company (Scale AI, Nuro/Cruise data & perception), or a massive‑scale data platform (ByteDance, Google).

• Base Cash Salary: $300,000 - $500,000 USD + Equity: Generous founding‑team equity package
• Benefits: Premium health, dental, and vision insurance, 401(k), and comprehensive Silicon Valley tech perks