AI Supply Chain & Data Ops Intern
Build and scale the data pipelines and MLOps infrastructure that power Cehpoint’s AI security products.
Role Overview & Operational Scope
You will own the data supply chain that feeds Cehpoint’s AI-driven threat detection and vulnerability-assessment systems — from ingestion and governance through model training and production serving. Your work ensures data reliability, reproducibility, and low-latency delivery so every model iteration is backed by trusted, auditable pipelines.
Key Responsibilities & Production Deliverables
- Design and maintain ETL/ELT pipelines for ingesting, validating, and transforming threat intelligence feeds, container images, and network telemetry into feature stores.
- Orchestrate data workflows using Apache Airflow or Prefect, implementing SLA monitoring, retries, dead-letter handling, and alerting on pipeline failures.
- Containerize training and inference workloads with Docker, package them on Kubernetes, and manage resource allocation across GPU/CPU nodes.
- Version datasets, schemas, and model artifacts via DVC or MLflow; ensure lineage and auditability for compliance (ISO 27001, SOC 2, DPDP Act 2023).
- Implement CI/CD for data and ML pipelines, including automated tests, linting, schema validation, and rollback strategies.
- Collaborate with the security engineering team to harden pipeline environments — secrets management, least-privilege IAM, network segmentation, and encrypted data-at-rest/in-transit.
Mandatory Foundational Knowledge
- Relational and NoSQL data modeling, normalization/denormalization trade-offs, and query optimization fundamentals.
- MLOps lifecycle: data versioning, feature store concepts, model registry, drift detection, and continuous training patterns.
- Cloud-native architecture: managed compute, object storage, message queues, and serverless functions on AWS or GCP.
Mandatory Practical Skills & Architecture
- Python (pandas, PySpark or similar) and advanced SQL
- Docker & Kubernetes (kubectl, Helm basics)
- Apache Airflow or Prefect for workflow orchestration
- MLflow or DVC for experiment & artifact tracking
- AWS S3 / BigQuery or GCP Cloud Storage / BigQuery for data storage and querying
Problem Solving, Execution Rigor & Curiosity
- Hunts for silent data-quality failures — you instinctively add assertions and observability before they become incidents.
- Reads release notes and RFCs for every tool in the stack; you prototype new features before they hit production.
- Questions the 'default' pipeline design and pushes for cheaper, faster, or more resilient alternatives.
5-day live technical evaluation milestone
Over a 5-day milestone, you will build an end-to-end MLOps pipeline: ingest a synthetic threat-intelligence dataset (CSV + JSON), validate and transform it with Python/SQL, orchestrate the workflow with Airflow, track experiments and artifacts with MLflow, containerize the training script with Docker, and deploy a minimal inference endpoint on Kubernetes. The pipeline must include error handling, schema validation, a README with runbooks, and a short demo recording.
Institutional hiring protocol: candidates who clear resume screening take a live practical milestone of strictly 5 working days. Verifiable completion and code audit by your assigned engineering mentor is the sole prerequisite for the official offer letter.
Compensation, Total Rewards & Advancement
- Competitive annual stipend (₹12L – ₹20L) with performance-based acceleration.
- Remote-first setup — your workstation and home-office allowance covered.
- Generous GPU cloud credits for experimentation and personal projects.
- Direct mentorship from senior security engineers and ML architects.
- Path to full-time offer with equity considerations for top performers.
Dedicated inbox for this role
Questions, private repository links or portfolio references for this opening route straight to the engineering leads reviewing it.
ai-supply-chain-data-ops-intern-careers@cehpoint.co.in
Open mail client →