Full-Stack Engineer
Oct 2023 – Present Vividly · Remote
B2B SaaS platform for enterprise trade promotion management · 2,500+ professional users · $63M raised · 20-person engineering team
- Built and set the roadmap for an internal developer platform used by the entire engineering team, including during every on-call incident: a Python CLI, an MCP server (FastMCP) with AI agent support, and GitHub Actions CI/CD integrations. Includes a performance scanner integrating Sentry, GCP Cloud Trace, and Query Insights, PR health reporting delivered to Slack, and a Django public API code generator with Cursor AI agent support.
- Designed and shipped a full-stack AI Contract Generation Engine on Django: template analysis, field extraction, and a verification pipeline built on LLMs, with a React and TypeScript interface. Cut contract creation from roughly 2 hours to under 5 minutes, a 95% cut in manual effort for 2,500+ enterprise users. Also built AI Scenario Planning for trade promotion simulation.
- Designed and implemented an ERP Integration framework on Django and PostgreSQL, with URL allowlisting and SSRF guardrails derived from a written threat model. Built the Integrations Trigger Page (React, TypeScript, GraphQL) so business administrators configure and trigger third-party connections themselves, no engineering involvement needed.
- Architected the Promotion Email Generation system end to end on Django and Strawberry GraphQL: database schema, email templates backend, a SendPromotionEmail mutation, Celery async delivery, and the full React and TypeScript interface.
- Own the Rust forecast microservice (axum, Rayon, SQLx), updated to respect user-configurable lift preferences across millions of rows. Built the companion Forecast Version Comparison UI (React, TypeScript, GraphQL) with weekly drill-down.
- Built the Budget 2.0 filter system (React, TypeScript, GraphQL): fund type filters, a date-pivot table, and clear-filters UX. Optimized the underlying GraphQL query performance in Django, and used BigQuery to fix budgeting and forecast revenue alignment discrepancies.
- Built the backend validation framework for the Promotions model in Django and PostgreSQL, covering 7+ field validators. Implemented the Promotion Submission and Approval workflow and Advanced Promotion Duplication.
- Built the IFM (Indirect Financial Model) Updater as a scheduled GCP Cloud Run job in Python for automated financial recalculations, and restructured the underlying IFM data model. Added ROI, revenue, and margin columns to promotion tables.
- Own around 10 production API endpoints across REST and GraphQL. The three highest-traffic hold p95 latency under 100ms; cut query latency on the biggest datasets from ~3s to 500ms (6x) by profiling with GCP Query Insights and cleaning up N+1 patterns.
- Wrote tests at every layer: Vitest for the frontend with a configurable mock data factory, pytest for the backend with coverage-exclusion paths configured so the number means something, and integration tests for the Rust services. Built a custom ESLint plugin for test conventions and maintained Storybook.
- Led a live migration of a production financial calculation system without downtime: ran the old and new versions in parallel with automated output comparison, then flipped traffic over with feature flags once the numbers matched. That approach is now the team's playbook for anything similarly risky.
- Led project teams of 4-5 engineers and worked directly with product managers, designers, and business analysts. Interviewed, helped hire, and onboarded 5 engineers, held regular 1:1s.