Integrations Engineer — ComplyDP
ComplyDP is a DPDP compliance platform. We connect to enterprise systems read-only, discover where personal data lives, and produce audit-grade records of processing, without ever copying client data. Backed by NVIDIA Inception, Google Cloud for Startups, T-Hub and DST. Clients include listed manufacturing groups, global research firms and regulated financial entities.
You will own the connector layer. It is the part of the product that reaches into a client's estate, reads structure and metadata, and returns an inventory. Zero replication is architectural, not aspirational: connectors operate on metadata and counts only.
Indian enterprise estates are not cloud-native. A large share of what we need to reach sits on-prem, including SQL Server, Oracle, file shares, Active Directory and legacy line-of-business systems. Reaching it, outbound-only and without asking a client's network team to open anything inward, is a defining part of this role.
What you'll build
Read-only connectors into cloud storage, SQL/NoSQL databases and SaaS APIs across enterprise estates
On-prem and legacy connectors: agents deployed inside client networks, operating outbound-only, with no inbound access to the client estate
Schema and metadata introspection that identifies personal-data columns without reading records
Detection of India-specific identifiers (Aadhaar, PAN, GSTIN) via pattern and checksum logic
Vendor and data-exchange mapping: what personal data enters and leaves each third-party processor, derived from configuration and code rather than self-declaration




Repository configuration analysis to surface undeclared flows
You should have
3 to 7 years backend or integration engineering
Robust Python: asyncio, FastAPI, SQLAlchemy (Inspector, not just ORM)
Production OAuth 2.0: authorization code and client credentials, token refresh, webhook signature verification
Database metadata introspection: information_schema, pg_catalog, read-only role provisioning
REST API fluency: pagination, rate limits, backoff, idempotency
Experience deploying software inside a customer's network: containerised agents or collectors, outbound-only connectivity, egress proxies, and the constraints enterprise IT and network security will actually impose
Comfort with on-prem sources (SQL Server, Oracle, SMB/NFS file shares, LDAP/Active Directory) alongside cloud and SaaS
Azure or AWS, secrets management, managed identity. Docker; working Linux and Windows networking
Judgment about credentials. You will hold read-only access into regulated enterprises. Security instinct is the non-negotiable
Bonus: prior connector work at an iPaaS, ETL, data-discovery or enterprise security vendor. Exposure to privacy or security tooling.
This is not a data engineering role. No pipelines, no warehouses, no Spark. The problem is breadth of systems and correctness of access, not volume of data.
Apply with a short note on a connector you shipped: what system it read from, how you authenticated, and how you got schema or metadata out without pulling the underlying records. If you've deployed anything inside a customer's own network, tell us how you handled connectivity.

📌 Integration Engineer (Khairahani)
🏢 ComplyDP
📍 Khairahani