The Opportunity
The intelligence in our Jobsite Intelligence Tools lives in the models you will build. Our hardware generates rich, novel data streams from custom-designed RF sensors, UWB radar, and other sensing modalities, data that has no off-the-shelf model waiting for it. Your job is to develop the machine learning systems that turn those raw signals into actionable insight: What material is behind this wall? How far away is that rebar? What changed since the last scan? This is not a role for someone who fine-tunes pretrained vision models and calls it a day. We need someone who is comfortable working with unconventional sensor data, building training pipelines largely from scratch, and deploying models onto constrained edge hardware. If that challenge excites you, we want to talk.

What You’ll Do
Model Development & Sensor Data Analysis

• Design, train, and evaluate machine learning models to extract meaningful insight from RF, UWB radar, and other sensor data streams unique to the as-built environment.

• Develop signal processing and feature engineering pipelines that transform raw, high-dimensional sensor data into representations suitable for learning.

• Build classification, regression, segmentation, and anomaly detection models appropriate to jobsite sensing tasks (material identification, distance estimation, structural characterization, change detection).

• Work closely with RF and hardware engineers to understand sensor characteristics, noise profiles, and data artifacts — and turn that knowledge into better models.

• Iterate rapidly on model architectures; know when a simple model beats a complex one and when it doesn’t.

Data Infrastructure & Labeling

• Build and maintain data collection pipelines from prototype hardware in lab and field settings.

• Design and manage datasets for novel sensing modalities where labeled data does not yet exist — including defining labeling strategies and ground truth collection methods.

• Develop tools and workflows for data annotation, curation, versioning, and quality control.

• Identify data gaps and work with the hardware team to instrument collection campaigns that address them.

Edge Deployment & Optimization

• Optimize trained models for deployment on resource-constrained edge hardware: quantization, pruning, knowledge distillation, and operator fusion.

• Work with embedded software engineers to integrate models into the on-device inference pipeline using frameworks such as TensorFlow Lite, ONNX Runtime, or vendor-specific SDKs.

• Profile and benchmark inference latency, memory footprint, and power consumption on target hardware.

• Make principled tradeoffs between model accuracy and the real constraints of running at the edge in a battery-powered handheld tool.

Experimentation & Research

• Stay current with relevant ML research — signal processing, radar ML, sensor fusion, and efficient on-device inference — and apply it where it matters.

• Design rigorous experiments; track hypotheses, results, and learnings in a reproducible, shareable way.

• Contribute to internal documentation of model architectures, dataset decisions, and deployment configurations.

• Evaluate and integrate third-party models, pretrained weights, or open-source components where they provide genuine leverage.

 
What We’re Looking For
Required Experience

• 2–5 years of professional experience developing and deploying machine learning models.

• Strong Python fluency; deep familiarity with PyTorch and/or TensorFlow.

• Experience working with non-image, non-text data — time series, signal data, sensor streams, or similar — is essential. 

• Demonstrated ability to build training data pipelines and datasets, including in domains where labeled data is scarce or must be created from scratch.

• Solid grounding in ML fundamentals: loss functions, regularization, evaluation metrics, cross-validation, and avoiding the many ways models can fool you.

Technical Depth

• Experience with model compression and on-device deployment (TFLite, ONNX, CoreML, TensorRT, or similar).

• Comfort with data versioning and experiment tracking tools (DVC, MLflow, Weights & Biases, or equivalent).

• Proficiency with scientific Python stack: NumPy, SciPy, pandas, matplotlib.

• Familiarity with Linux-based development environments and embedded Linux targets.

Startup DNA

• Comfortable building from scratch — no existing datasets, no existing model baselines, no existing pipelines on day one.

• Intellectually curious about the physical world; you find it interesting that RF signals can reveal what’s hidden behind a wall.

• Pragmatic problem solver: you can distinguish between a research problem and an engineering problem, and you know how to make progress on both.

• Clear communicator who can explain model behavior, limitations, and confidence to hardware engineers, product managers, and field users who are not ML practitioners.

 
Nice to Have

• Background in physics, electrical engineering, or a related technical discipline — the ability to reason about what the sensor is physically measuring matters here.

• Experience with few-shot learning, self-supervised learning, or domain adaptation in low-data regimes.

• Familiarity with C/C++ for performance-critical inference code or embedded integration.

• Experience deploying models to microcontrollers or DSPs (Cortex-M, RISC-V, Hexagon DSP, or similar).

• Prior work in construction tech, industrial sensing, geophysics, or physical inspection applications.

 
Compensation & Benefits

We offer a package built for someone energized by the problem, not just the paycheck:

• Base salary that is intentionally below market — we are a startup and we are transparent about that tradeoff.

• Equity that reflects an early, meaningful technical contribution: you are not joining a mature ML team; you are building something that does not yet exist, and your grant will reflect that.

• Full health and dental coverage.

• Flexible PTO and a culture that actually respects it.

• Access to the hardware, compute, and lab resources you need to do the work.

• Direct exposure to the full product stack — from raw sensor physics to shipped intelligence.

 
How to Apply
Send your resume and a brief note on why this role excites you to careers@resolvtools.com. We review applications on a rolling basis and move quickly with strong candidates.
 
Resolv Tools is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.