About the role
As an ML Engineer II, you will contribute to building and scaling InMobi's ML platform by developing reliable ML pipelines, training workflows, and serving systems. You will work closely with senior engineers to operationalise ML models and improve developer productivity for data scientists and ML engineers.
Responsibilities:
- Build and enhance components of a self-serve ML platform supporting training, evaluation, and inference workflows.
- Develop and deploy ML models using PyTorch, TensorFlow, or LightGBM, supporting the end-to-end ML lifecycle from experimentation to production.
- Build and operate scalable, low-latency models serving pipelines using Triton, TorchServe, or TensorFlow Serving.
- Implement MLOps workflows using MLflow for experiment tracking and Airflow (or similar) for orchestration.
- Work with Kubernetes-based ML workloads, including resource configuration, autoscaling, and reliability improvements.
- Add observability (metrics, logs, dashboards) to monitor training and inference systems and improve service stability.
Requirements:
- 3-9 years of experience building and deploying production ML systems.
- Strong Python programming.
- Hands-on experience with PyTorch / TensorFlow.
- Familiarity with ML training, evaluation, and inference workflows.
- Basic experience with distributed systems or ML workloads.
- Exposure to Kubernetes and cloud platforms (GCP preferred).
- Understanding of feature engineering and model monitoring concepts.
Good to Have:
- Experience with Ray (training or tuning).
- Knowledge of feature stores and online/offline pipelines.
- Familiarity with observability tools (Prometheus, Grafana).
- Exposure to Kafka, Spark, or Flink.