Software Engineer: Perception (II,III)
About Lodestar
Lodestar is the software-first space prime for the hardware abundant era. We believe the next generation of space systems will be defined not by launch, fuel, and batteries, but by software, compute, and autonomous effects.
Space is becoming an increasingly contested warfighting domain, with growing threats to the satellites and critical infrastructure underpinning modern life. Maintaining Western strategic advantage will require a new generation of space capabilities built for the speed, complexity and scale of future conflict.
Lodestar is addressing this challenge by building superintelligent spacecraft capable of operating autonomously in orbit. Combining our capabilities in advanced adversarial tracking & characterization, onboard edge compute and autonomous effect delivery, our technology enables spacecraft to execute complex rendezvous and proximity operations (RPO), detect and classify objects, predict intent and respond to orbital threats - even in denied environments. By integrating sensing, reasoning and action into a single autonomous system, we aim to deliver decisive situational awareness and space superiority for defense and commercial operators.
Backed by elite deep-tech venture capital, international space agencies and government stakeholders, Lodestar is rapidly expanding its footprint across the U.S. Space Force, the Department of War, and major commercial satellite operators. We are a lean, high-autonomy team of engineers and operators building the space defense infrastructure of tomorrow.
Join us.
About the Job
As a Software Engineer II/III - Perception, you'll combine classical and modern CV and ML methods to build the custom perception system at the core of Lodestar’s autonomy software suite, MITHRIL. You'll own, build and refine perception models and work within a team to deliver systems to detect, classify and understand unknown space targets in real time. Challenges like harsh sunlight, reflective surfaces, extreme contrast and little labeled data rarely have off-the-shelf answers, so much of the work is understanding why an approach fails and designing something better. This is a mid-career role for someone with research depth. You should be able to explain and defend your work (architecture, loss, metrics, data), not just report benchmark results.
We proudly have an “extreme ownership” oriented engineering culture. You’ll get real ownership early, with support to grow into it.
What You’ll Do
- Research, design and evaluate perception methods at both model and system level for detecting, classifying and characterizing unknown space targets, from literature review to working model
- Develop methods for estimating the relative pose and inferred capabilities of non-cooperative space targets under harsh lighting, high contrast and limited data
- Design and improve training strategies and associated systems where real imagery is scarce: synthetic data generation, domain adaptation, and measuring the sim-to-real gap
- Run rigorous experiments: ablations, failure analysis, and metrics that reflect mission needs, not generic benchmarks
- Integrate perception models into mission simulation environments, real-time autonomy pipelines and compute-limited edge nodes working with the autonomy and software teams
- Contribute to physically realistic sensor models (optical, thermal, radar, lidar) to improve perception fidelity in mission scenarios
- Maintain and extend our perception tooling, model training infrastructure, and evaluation frameworks
Basic Qualifications
- Strong ML fundamentals: you can explain why an architecture, loss function or training setup suits a problem, not just that it scored well
- Hands-on computer vision experience, both classical (geometry, feature-based methods) and deep learning
- Degree in Computer Science, Electrical Engineering, Physics, Aerospace, Robotics, AI/ML or a related field, or equivalent experience
- 2–5 years in ML, computer vision or robotics. Graduate research (Master’s or PhD) counts
- Have taken a model from problem definition through research and design, training and evaluation, and can walk through its decisions, trade-offs and failures in depth
- Evidence of going beyond applying existing methods: a new method, a publication, a significant modification of an existing model, or a substantial open-source contribution
- Exposure to sensor data from electro-optical, infrared, radar, RF and/or lidar sensors
- Full proficiency in Python
Preferred Skills & Experience
- 6-DoF pose estimation, or tracking and state estimation
- 3D reconstruction (multi-view geometry, neural rendering, Gaussian splatting)
- Space or space-adjacent experience: space domain awareness (ground- or space-based), space imagery, astronomy or low-light imaging, rendezvous and proximity operations, or spacecraft GNC
- Synthetic data generation and sim-to-real transfer
- Sensor physics, e.g. noise models, radiometry, thermal signatures
- Publications at computer vision, ML, robotics or aerospace venues (e.g. CVPR, ICCV, ECCV, NeurIPS, ICRA, AIAA, AMOS)
- Optimizing models for edge hardware (TensorRT, CUDA, quantization)
- Working proficiency in C++
- Linux, Git, CI/CD and Docker
ITAR Requirements
To conform to U.S. Government space technology export regulations, including the International Traffic in Arms Regulations (ITAR) you must be a U.S. citizen, lawful permanent resident of the U.S., protected individual as defined by 8 U.S.C. 1324b(a)(3), or eligible to obtain the required authorizations from the U.S. Department of State. Learn more about the ITAR here.
Compensation & Benefits
- Pay Range for this role:
- (P2) Developing: $115,000 - $174,000 / year
- (P3) Career: $136,000 - $200,000 / year
- Meaningful equity incentives as part of our employee option pool
- Flexible PTO with generous paid vacation, holidays, and sick leave
- Comprehensive medical, dental & vision coverage
- 401(k) retirement plan with company match
- Los Angeles
Additional Information
Compensation bands are determined by role, level, location, and alignment with market data. Individual level and base pay is determined on a case-by-case basis and may vary based on job-related skills, education, experience, technical capabilities and internal equity. In addition to base salary, for full-time hires, you may also be eligible for long-term incentives, in the form of stock options , and access to medical, vision and dental coverage, as well as access to a 401(k) retirement plan.
Lodestar is an Equal Opportunity Employer; employment with Lodestar is governed on the basis of merit, competence and qualifications and will not be influenced in any manner by race, color, religion, gender, national origin/ethnicity, veteran status, disability status, age, sexual orientation, gender identity, marital status, mental or physical disability or any other legally protected status.
Create a Job Alert
Interested in building your career at Lodestar? Get future opportunities sent straight to your email.
Apply for this job
*
indicates a required field