Data Scientist – Edge AI & Embedded Systems

Vantor•Reston, VA

13d•Onsite

About The Position

Vantor is seeking a Data Scientist to join a small team of experts focused on bringing modern AI/ML capabilities to the tactical edge. This role involves taking a project from concept to completion, including identifying use cases, sourcing and curating data, recommending models and applications, training and fine-tuning models, deploying them onto embedded hardware in air-gapped environments, and ensuring their usability. The ideal candidate is curious, detail-oriented, enjoys working with data, can analyze model performance, troubleshoot Linux drivers on single-board computers, and effectively communicate complex concepts to non-technical stakeholders. This is an on-site position in Reston, VA, working closely with the customer, mission operators, and requirements owners.

Requirements

Bachelor's degree or higher in an applicable technical degree, such as data science, computer science, computer engineering, statistics, or electrical engineering.
At least five years of relevant experience.
U.S. Citizen with a TS/SCI clearance and CI poly.
Solid understanding of AI/ML algorithms, including computer vision models and LLMs, and hands-on experience training and evaluating models using frameworks such as PyTorch or TensorFlow.
Proficiency in Python and its data science ecosystem (e.g., NumPy, pandas, scikit-learn), plus Bash scripting.
Experience configuring and building applications in Linux and deploying offline-native applications with Docker.

Nice To Haves

Previous experience working on-site to support the DoD or the IC.
Experience with data preparation, including cleaning, labeling, augmentation, and managing datasets.
Awareness of the current open-source AI/ML landscape and familiarity with GitHub, GitLab, Hugging Face, and Kaggle.
Contributions to open-source projects.
Experience optimizing models for edge deployment (quantization, pruning, TensorRT, ONNX Runtime, llama.cpp).
Familiarity with embedded system design and troubleshooting low-level Linux issues.
Understanding of common hardware architectures, including x86, ARM, SoC, and RISC-V.
Experience with any Single Board Computer (SBC), including Raspberry Pi, Arduino, BeagleBoard, NVIDIA Jetson, etc.
Experience with lower-level languages such as C, C++, or Java.
Experience analyzing RF or signal data; familiarity with Software Defined Radios (SDRs).
Understanding of modern digital communication systems, including RF, cellular, and WiFi.
Experience with MLOps practices such as model versioning and experiment tracking in offline environments.
Ability to communicate complex technical topics to a range of audiences, including technical and non-technical leaders.

Responsibilities

Source, curate, clean, and label datasets from open-source repositories (GitHub, GitLab, Hugging Face, Kaggle) and operational sensor data.
Train, fine-tune, and evaluate AI/ML models, including computer vision models and LLMs, for specific mission use cases.
Design evaluation methods and metrics to measure model accuracy, latency, and reliability under real-world conditions.
Optimize models for resource-constrained hardware using techniques such as quantization, pruning, and conversion to edge runtimes.
Build and deploy offline-native data and ML pipelines in containerized environments using Docker.
Configure and troubleshoot Linux-based embedded systems, including compatibility between specific kernels and hardware drivers.
Explore and analyze data from digital communication systems, including RF, cellular, and WiFi sources.
Evaluate open-source AI software and models to identify solutions that meet the customer's operational requirements.
Engage with customers to understand expectations and ensure delivered products meet operational requirements.
Produce supporting documentation, including model performance reports and inputs for user manuals.