ABOUT THE COMPANY

Our client is a private technology company that builds digital infrastructure for critical physical assets — software systems for monitoring, control, and lifecycle management of high-value platforms across maritime, aerospace, ground, and subsurface domains. Its products serve defence and industrial clients where reliability, security, and real-time performance are non-negotiable.

ABOUT THE ROLE

You will own production data pipelines deployed inside customer environments in New Delhi — ingesting, normalising, and serving heterogeneous sensor, geospatial, and tabular feeds into a unified data layer that mission analysts depend on. The role is forward-deployed: most of your week is spent on customer premises, often in air-gapped or restricted-network conditions, working shoulder-to-shoulder with operators to ship data products that move the mission needle.

KEY RESPONSIBILITIES

— Design and operate Airflow-orchestrated pipelines ingesting high-volume sensor, satellite, and third-party feeds; own SLAs end-to-end.

— Build Spark workloads for batch enrichment, geospatial joins, and large-scale aggregation across time-series and tabular layers.

— Model and tune the storage tier on PostgreSQL with TimescaleDB and PostGIS extensions; manage bulk-loading via COPY pipelines.

— Containerise the stack with Docker / Compose; package air-gapped delivery bundles for installation at customer sites without internet egress.

— Sit alongside customer analysts — translate operational questions into data contracts, dashboards, and queryable surfaces.

— Diagnose data-quality and performance issues under production conditions; restore service quickly and write the postmortem honestly.

— Document lineage, schemas, and runbooks so customer engineering teams can operate the stack between visits.

QUALIFICATIONS & SKILLS

EXPERIENCE: Minimum 6 years in production data engineering, including ownership of pipelines running at customer or production scale.

ORCHESTRATION: Apache Airflow at depth — DAG design, sensors, custom operators, backfills, and operating Airflow itself in production.

COMPUTE: Apache Spark for batch — Spark SQL, partitioning, shuffle and memory tuning. Familiarity with columnar formats (Parquet, Arrow).

STORAGE: Deep PostgreSQL proficiency. Working knowledge of TimescaleDB (time-series), PostGIS (spatial), and DuckDB for in-process analytics.

CONTAINERS: Docker and Compose; comfortable assembling and shipping air-gapped delivery bundles. Nginx as reverse proxy / static serving.

LANGUAGES: Production Python (FastAPI/Pydantic a plus). TypeScript / Node for service code is a strong plus.

DOMAIN: Geospatial and time-series data — GeoJSON, MGRS, satellite ephemeris (SGP4), sensor telemetry. S3-compatible object storage in the loop.

DISPOSITION: Delhi-based or willing to relocate. Customer-facing posture, security awareness, and comfort operating from secured customer premises.