Founding Machine Learning Engineer
(Magnetic Navigation)
Founding Machine Learning Engineer
(Magnetic Navigation)
Location: Bengaluru, Karnataka, India
Job Type: Full-time
Work Mode: On-site and Occasional Field Testing
Location: Bengaluru, Karnataka, India
Job Type: Full-time
Work Mode: On-site and Occasional Field Testing
About
About
Dirac Labs is building diamond NV-center magnetometers and related quantum sensors that unlock precise navigation when GPS is unavailable. Our work is supported by NASA, NOAA, Emergent Ventures, and the USISTEF. Our mission is simple: provide trustworthy navigation in every location without worrying about spoofing or jamming.
Dirac Labs is building diamond NV-center magnetometers and related quantum sensors that unlock precise navigation when GPS is unavailable. Our work is supported by NASA, NOAA, Emergent Ventures, and the USISTEF. Our mission is simple: provide trustworthy navigation in every location without worrying about spoofing or jamming.
Job Description
Job Description
As the Founding ML Engineer, you will develop core algorithms and models that transform raw magnetic and inertial signals into accurate, real-time location estimates. You will own the end-to-end ML stack, from denoising and map-matching to embedded deployment and work closely with hardware and field teams to close the loop between models and the real world.
As the Founding ML Engineer, you will develop core algorithms and models that transform raw magnetic and inertial signals into accurate, real-time location estimates. You will own the end-to-end ML stack, from denoising and map-matching to embedded deployment and work closely with hardware and field teams to close the loop between models and the real world.
Key Responsibilites
Key Responsibilites
Magnetic Signal Processing:
Design and train denoising models across diverse environments and platforms.
Map-Matching:
Build pipelines to align live magnetic observations with prior maps for pose and position inference.
Unified Modeling:
Architect a model that fuses temporal magnetic data, inertial cues, and contextual priors with uncertainty bounds.
Data Tooling:
Create pipelines for collection, labeling, augmentation, and synthetic generation of trajectories.
Evaluation & Deployment:
Define metrics, run replays/A-B tests, and deploy on embedded devices.
Collaboration & Documentation:
Work with hardware/firmware teams and publish clear specs and studies.
Magnetic Signal Processing:
Design and train denoising models across diverse environments and platforms.
Map-Matching:
Build pipelines to align live magnetic observations with prior maps for pose and position inference.
Unified Modeling:
Architect a model that fuses temporal magnetic data, inertial cues, and contextual priors with uncertainty bounds.
Data Tooling:
Create pipelines for collection, labeling, augmentation, and synthetic generation of trajectories.
Evaluation & Deployment:
Define metrics, run replays/A-B tests, and deploy on embedded devices.
Collaboration & Documentation:
Work with hardware/firmware teams and publish clear specs and studies.
Required Qualifications
Required Qualifications
2+ years in applied ML or strong research background with shipped systems
Expertise in sequence modeling or sensor fusion (e.g., Transformers, SSMs, Kalman variants, diffusion-style denoisers)
Experience building real-time inference systems (quantization, pruning, profiling)
Strong skills in Python and one systems language (preferably C++)
Practical experience with data collection, cleaning, and stress-testing
Clear writing and ability to turn experiments into decisions
2+ years in applied ML or strong research background with shipped systems
Expertise in sequence modeling or sensor fusion (e.g., Transformers, SSMs, Kalman variants, diffusion-style denoisers)
Experience building real-time inference systems (quantization, pruning, profiling)
Strong skills in Python and one systems language (preferably C++)
Practical experience with data collection, cleaning, and stress-testing
Clear writing and ability to turn experiments into decisions
Preferred Qualifications
Preferred Qualifications
Background in SLAM, map-matching, or PNT
Edge/embedded ML deployment (Jetson, Qualcomm, ARM)
Experience with time-series foundation models, self-supervised or contrastive learning
Synthetic data, domain randomization, or physics-informed learning
Familiarity with inertial sensors, magnetometers, or geophysics data
Background in SLAM, map-matching, or PNT
Edge/embedded ML deployment (Jetson, Qualcomm, ARM)
Experience with time-series foundation models, self-supervised or contrastive learning
Synthetic data, domain randomization, or physics-informed learning
Familiarity with inertial sensors, magnetometers, or geophysics data
Our expectations
Our expectations
90 days: a reproducible denoising baseline with measurable lift on our replay datasets.
6 months: a closed-loop prototype that locates within target error bounds on real routes, with on-device inference at required latency.
12 months: a unified model and map-matching stack that generalizes across cities and facilities with robust spoofing and jamming resilience.
90 days: a reproducible denoising baseline with measurable lift on our replay datasets.
6 months: a closed-loop prototype that locates within target error bounds on real routes, with on-device inference at required latency.
12 months: a unified model and map-matching stack that generalizes across cities and facilities with robust spoofing and jamming resilience.
How to apply
How to apply
Email your CV and a brief project portfolio to aishwarya@diraclabs.com with the subject “Founding ML Engineer.”
Include a cover letter on a sensor-fusion system you built or the hardest bug you faced, and how you resolved it.
Email your CV and a brief project portfolio to aishwarya@diraclabs.com with the subject “Founding ML Engineer.”
Include a cover letter on a sensor-fusion system you built or the hardest bug you faced, and how you resolved it.
© 2025 Dirac Labs. All rights reserved.
© 2025 Dirac Labs. All rights reserved.
© 2025 Dirac Labs. All rights reserved.