AI Engineer, Time-Series Signal Processing
BrightAI is a high-growth Physical AI company transforming how businesses interact with the physical world through intelligent automation. Our platform processes visual, spatial, and temporal data from billions of real-world events—captured through edge devices, mobile sensors, and large-scale cloud infrastructure—to deliver intelligent, real-time decisions.
We are now hiring an AI Engineer – Time-Series Signal Processing to lead the development of AI/ML solutions built on high-frequency multi-modal sensor data. This is a critical role focused on modeling and understanding time-series signals coming from IoT devices equipped with various sensors (IMU, acoustic, pressure, temperature, etc.) that drive intelligent automation across physical infrastructure systems.
You'll work on building cutting-edge real-time AI models that process noisy, high-throughput data streams and extract meaningful insights for real-world decision-making—at both the edge and cloud scale.
Responsibilities
- Design and implement real-time signal processing and ML pipelines for multi-modal time-series data such as those acquired from IMUs, microphones, pressure or force sensors, ultrasonic transducers, and similar sensor sources.
- Develop and deploy ML models for time-series classification, prediction, anomaly detection, activity recognition, condition monitoring and pattern analysis.
- Lead research and implementation of RNN-based architectures (especially LSTMs and their variants) as well as temporal transformer models as needed.
- Build and tune classical and tree-based ML models (XGBoost, LightGBM, Random Forests, and other gradient-boosted ensembles) for time-series tasks, including feature engineering and model interpretability (e.g., SHAP).
- Work with SCADA systems and industrial telemetry data—ingesting and modeling high-frequency, multi-channel operational data streams from physical assets.
- Collaborate with hardware, embedded, and product teams to integrate models into edge devices and IoT platforms.
- Drive experimentation and optimization of signal-processing techniques (e.g., filtering, feature extraction, event detection) to enhance model input quality.
- Design and maintain scalable workflows for ingesting, labeling, training, and evaluating multi-channel time-series datasets.
- Stay current with advances in time-series modeling, signal processing, and real-time inference, and incorporate them into product roadmaps.
- Ensure model robustness, performance, and reliability in production environments, including edge deployments.
Educational Background
- Degree in Electrical Engineering, Computer Science, or a related field, with a strong focus on signal processing, time-series analysis, and machine learning.
- Strong academic or industry track record in time-series modeling, signal processing, or real-time AI systems.
Required Skills & Expertise
- 2+ years of experience developing signal processing and ML solutions for time-series sensor data. Track record of bringing at least one ML solution to market.
- Deep understanding of digital signal processing (DSP) methods: filtering, sampling, windowing, FFT, feature extraction, etc.
- Hands-on experience with RNNs (especially LSTMs/GRUs) and/or temporal convolutional networks for time-series modeling.
- Proficiency with tree-based and gradient-boosting models (XGBoost, LightGBM, Random Forests) applied to time-series and sensor data, including hyperparameter tuning and explainability.
- Experience working with SCADA systems and industrial telemetry data (high-frequency sensor feeds, time-stamped operational data, multi-channel ingestion from physical assets).
- Proven experience with time-series data from physical sensors such as IMUs, microphones, vibration or pressure sensors.
- Strong coding skills in Python and fluency with ML/DL frameworks (e.g., PyTorch, TensorFlow, Keras).
- Experience in optimizing and deploying models in real-time or near-real-time environments, including edge devices or resource-constrained embedded systems.
- Fluency with best practices in data labeling, augmentation, and evaluation for time-series tasks.
- Excellent problem-solving and collaboration skills with the ability to work across teams.
- Strong communication skills with the ability to convey findings and recommendations to internal and external stakeholders.
Bonus Qualifications
- Experience building end-to-end AI systems for structural health monitoring, condition monitoring, anomaly detection, activity recognition, or motion tracking.
- Experience with predictive maintenance on industrial equipment using SCADA/telemetry data.
- Familiarity with experiment tracking and model lifecycle tooling (e.g., MLflow, DVC).
- Exposure to streaming/online inference patterns (e.g., EWMA normalization, windowed feature extraction on live data).
- Proficiency in embedded software or deploying models to constrained environments (e.g., using TFLite, ONNX, or custom firmware).
- Familiarity with containerized workflows and Linux-based development environments.
- Experience with Agile workflows and tools such as JIRA, Git, and CI/CD pipelines.
- Prior work in startup or high-pace teams with experience in building real-time systems from the ground up.