Software Engineer - Localization, State Estimation & Prediction

Lodestar•San Francisco, CA

14d•$120,000 - $170,000

About The Position

At Lodestar, as a Software Engineer - Localization, State Estimation & Prediction , you’ll be leading the development of our state estimation and prediction models at the core of Lodestar’s flagship product, MITHRIL . You’ll be focusing on researching and developing algorithms that fuse probabilistic estimation with machine learning to track targets, predict trajectories and assess intent. We proudly have an "extreme ownership" oriented engineering culture.

Requirements

Bachelor’s or Master’s degree in Computer Science, Aerospace, Robotics, Applied Mathematics, a related field, or equivalent experience
2+ years of distinguished industry experience in state estimation, aerospace, robotics, or a related field
Strong proficiency in C++ and Python
Demonstrated experience with probabilistic state estimation (Kalman filters, particle filters, Bayesian inference)
Experience with deep learning frameworks (PyTorch, TensorFlow) for sequence modeling (seq2seq, RNNs, LSTMs, Transformers)
Familiarity with trajectory modeling, multi-body dynamics, or orbital mechanics

Nice To Haves

Track record implementing state estimation pipelines in real-time or safety-critical systems
Experience deploying sequence-to-sequence models for forecasting in dynamic environments
Familiarity with sensor fusion techniques (e.g., combining radar, optical, and inertial data)
Proficiency with GPU acceleration (CUDA, TensorRT) for neural network training and inference
Experience with Linux, Git, and CI/CD pipelines
Comfortable with containerization tools such as Docker and Kubernetes
Familiarity with real-time systems, multi-threading, and performance optimization
Strong understanding of networking and communication protocols for distributed autonomy systems

Responsibilities

Design and implement Lodestar’s core state estimation and prediction architecture for autonomous spacecraft operations
Research and implement novel estimation and prediction algorithms - everything from lit. review, through training and tuning, to optimisation and deployment
Develop classical and neural estimators to track the current state and trajectory of multiple dynamic targets in real time
Create neural models to forecast future trajectories and behavioral patterns of targets
Implement intent inference models to identify actions and dynamically rank threat levels
Integrate state estimation and prediction models into mission simulation environments and autonomy decision systems
Collaborate with cross-functional teams including, perception, and on-board autonomy to ensure prediction fidelity, robustness, and scalability across missions