Position OverviewAllen Control Systems is seeking a Machine Learning (ML) Data Specialist to own the flow data through our computer vision and machine learning pipelines. You will be the steward of dataset quality and labeling throughput: overseeing the flow of incoming video and label data, reviewing outputs for correctness and quality, and surfacing distribution gaps and procedural issues to the ML Platform team. This will be a hybrid role out of our Austin, TX office.
What You'll Do:- Own end-to-end data and label flow for ACS ML training and testing, from raw inputs through label review and final ingest into training and testing datasets.
- Review labeled data for quality, correctness, and conformance to spec.
- Manage data routed to third-party labelers: selecting and prioritizing batches; tracking throughput and labeler quality.
- Audit incoming data from internal and third-party sources for coverage and balance; flag procedural issues with collection (e.g., over-collection of certain conditions, gaps in edge cases).
- Partner with the ML Platform team to provide insight into dataset composition, distribution gaps, and labeling coverage; help define and track the metrics that matter.
- Work in dashboards, spreadsheets, and Linux terminals to inspect data, run pre-built scripts, and maintain operational visibility into the pipeline.
- Document procedures and contribute to continuous improvement of data and labeling workflows.
Required Technical Skills:- Background: Bachelors Degree in a relevant field and 1+ years of relevant experience in data operations, dataset curation, annotation operations, QA or test engineering, video/imagery analysis, robotics or autonomy data ops, or a comparable role. New graduates with relevant project or internship experience also welcome.
- Eye for data quality: A careful, detail-oriented mindset for reviewing imagery and sensor data, spotting inconsistencies in labels, and reasoning about dataset coverage and edge cases.
- Operational ownership: Experience running a recurring operational process end-to-end - managing throughput, vendor coordination, QA pipelines, or similar - and driving improvements based on what you observe.
- Linux comfort: Basic command-line proficiency on Linux (navigating filesystems, running provided scripts, reading logs) and a willingness to grow that fluency on the job.
- Communication: Clear written communication to surface trends, raise issues, and align with internal engineering teams and external labeling vendors.
Preferred Technical Skills:- Domain exposure: Familiarity with computer vision concepts (object detection, tracking, segmentation) or prior work at an AV, robotics, drone, or other perception-heavy company.
- Light scripting: Basic Python, Bash, or SQL - or strong motivation to learn on the job.
What We Offer:- Health, Dental, Vision Insurance
#LI-AS1
