Deepak Goel

• Led Docker-based development of RL UI Gyms for Amazon, delivering production-quality environments for Zendesk, Jira, and Gmail; working on HubSpot and Doordash.
• Designed and implemented full-stack observability using OpenTelemetry, providing detailed frontend, backend, container, and node-level dashboards; instrumentation directly surfaced and resolved performance bottlenecks in Mira.
• Closed gaps in browser and Python auto-instrumentation by extending backend APIs, restoring missing semantic context, and resolving macOS vs Linux container discrepancies.
• Architected a deterministic test-data generation system with semantic inference, configurable distributions, and FK-safe incremental data generation across heterogeneous schemas.
• Built robust asynchronous data import pipelines handling circular dependencies, PII guardrails, constraint enforcement, and import history for large datasets.
• Designed Claude-based code-generation agents and a template repository to make AI-assisted development deterministic and aligned with quality standards.

• Led architecture of a cloud-native AppSec SaaS composed of 20+ microservices running on GKE (GCP) with focus on maintainability, observability, latency, reliability and throughput.
• Mentored engineers through design and code reviews; introduced platform standards, coding standards and review checklists that improved consistency and maintainability across services.
• Partnered with product managers and leadership to translate ambiguous requirements into concrete architectures and delivery plans.
• Designed and delivered a high-throughput ingestion pipeline using GCS, Dataflow, and Cloud Functions, increasing throughput by over 100x.
• Built a self-serve analytics layer powering dashboards and enabled product managers to create and iterate without engineering involvement.
• Designed and shipped LLM-assisted workflows for code-fix recommendations, PR review, architecture review, and test generation, with guardrails to ensure safety for production use.

• Architected a real-time analytics platform processing over 1M messages per second with sub-minute end-to-end latency.
• Designed efficient streaming, storage, and retrieval layers using Kafka, Spark, and Druid.
• Published streaming data into HDFS for batch analytics and ML workloads, enabling replay, backfills, and offline analysis.

• Designed multi-tenant SaaS solution that could elastically scale and was deployed on AWS.
• Built a configurable platform with custom fields/objects, forms, workflows, integrations, and a business rules execution engine to support divergent customer needs.
• Delivered platform features including SSO, fine-grained RBAC, notifications, scheduled jobs, and pipelines; enabled rapid onboarding across non-profits, manufacturing, and services.

Technologies: Java, React, Spring Boot, MySQL, AWS, SPA, CI/CD (GitHub Workflows), Terraform

• Architected backend capabilities for QuickBooks Online supporting 300K+ SMBs and millions of users, with strong reliability, performance, and production observability.
• Designed and implemented scalable financial workflows with strong consistency and auditability.
• Led cross-team initiatives across accountant workflows, core financial processing, authorization, and transaction categorization.
• Built a declarative, metadata-driven platform that reduced development time across product lines by standardizing business rules, validation, and persistence.

Technologies: J2EE, Spring MVC, JavaScript, Oracle, CI/CD (Jenkins)