Delhi, India
Backend engineer focused on building reliable, secure, data-heavy platforms and AI integrations. I care about clear interfaces, measurable outcomes, and systems that are easy to run anytime of the day. I’m comfortable owning problems end-to-end—design, implementation, rollout, and the monitoring that proves it works. What I bring: • Experience building multi-tenant platform tooling and access controls at scale. • A security-by-default mindset and the observability to back it up (dashboards, alerts, run books). • Strong collaboration across product, data, and SRE to keep delivery fast without trading off reliability. Select highlights: • Automated tenant provisioning with DRF + Boto3 (TDD-first) and supported the migration of 70+ legacy tenants. • Led a multi-DB redesign (Postgres + MongoDB) with Kafka/Mongo atomicity that reduced inconsistencies across 500+ measures and 700+ disease (valueset) datasets for 25+ customers. • Led a zero-downtime migration of 1TB+ container images and raised org-wide log availability to 99.97%. Currently exploring: • Robust patterns for LLM agents in production (safety, observability, cost controls). • Usable data governance and ACLs that make compliance easier, not harder. • Cost-aware architectures on GCP/AWS that scale without surprises. Core tools: Python, Java, Django REST Framework, LangGraph, GCP, AWS, BigQuery, Postgres, MongoDB, Kafka, Docker, Kubernetes, Jenkins, Helm, Prometheus, Grafana, Elasticsearch, Fluentd, GitLab, Celery. Checkout my experience below for specifics. If you’re building AI-driven platforms or scaling backend/infra with strong security and observability, I’d love to connect.
DEIB Analytics Pipeline (Great Place To Work insights) & Bryte Tenant Activation • Shipped a unified CLI to orchestrate 15+ components, cutting manual steps and errors. • Delivered GPTW & HCM correlation insights across 18 metrics to deepen DEIB analytics. • Implemented interface logic for Bryte Activation in Python and Java. DevOps and Security (Vulnerability Management Lead) • Ran secure data-handling sessions and automated policy enforcement via scripts. • Enforced PII policies in BigQuery with safe backups and deletion of legacy UserData. • Migrated 1TB+ Docker images from Artifactory/GCR to Google Artifact Registry with zero downtime. • Productionized services across 22 environments using custom in-house Quark configurations. • Remediated Jenkins pipeline vulnerabilities and upgraded Spark Operator CRDs for DSaaS projects. Tech: Python (FastAPI), Java (SpringBoot), LangGraph, GCP, BigQuery, Docker & Kubernetes, Jenkins, Spark Operator, CI/CD, Security & Compliance.
Identity & Access (User Management / ACL / Audit) • CLI scripts to migrate 50+ GB of user data across clients. • Asynchronous user-data export API with Celery + AWS S3 for 4 MongoDB collections. • Experience Builder notifications: 36+ custom account-event alerts used by multiple product backends. • Data Access Policies to create/assign policies for users; improved ACL-driven UX. • Eventing framework with Kafka for roles, groups, and data-access rules to downstream apps. Tenant Release/Admin Management (TRM/TAM) • DRF + Boto3 APIs to automate tenant creation and configuration (TDD-first). • Cross-team data collation and APIs to migrate 70+ legacy tenants. • Kafka consumer controller coordinating multiple microservices; consumes 5 topics for TRM/TAM. Logging & Observability (Org-wide) • Centralized logging with Elasticsearch + Fluentd; availability improved from 68.4% → 99.97% (S3 archival). • Monitoring with Prometheus + Grafana (Helm templates); 20+ dashboards for 50+ customers covering 26 core services each. • Prometheus exporters for 5+ internal metrics, including a scanner for 1,500+ GitLab repositories. • 3 Slack apps surfacing observability KPIs to leadership and client channels; Python backends on AWS Lambda. Measure Builder (Analytics) • Two custom operands for measures, built with Analytics team input. • Multi-DB redesign (Postgres + MongoDB) with Kafka + MongoDB atomicity; fewer inconsistencies across 500+ measures and 700+ disease datasets, impacting 25+ customers.
• Created WebAPI based Optical Character Recognition. • Development of server-side logic, ensuring high performance and responsiveness to requests from the front-end. • Used Image Processing to gain better accuracy.