GauriMarkandey.
Software engineer with 2+ years in production - building pixel-perfect UIs, design systems, and cloud-native backends that handle real load.
Work Experience
- Developed and enhanced Angular + TypeScript frontend on the NIH-funded Human Reference Atlas platform, integrating AWS S3 and building shared design system components and reusable libraries — ensuring WCAG accessibility and scalability, reducing frontend development overhead ~20%.
- Designed and implemented new user-facing interfaces in Angular and TypeScript, optimizing RxJS-driven data flows across critical user journeys and reducing redundant API requests ~20%.
- Owned production stability through monitoring, debugging, and resolving live issues — ensuring reliable releases with zero critical user-facing disruptions.
- Maintained CI/CD pipelines via GitHub Actions and leveraged AI-assisted developer tools to accelerate iteration and delivery, reducing build and release friction ~25%.
- Extended and migrated healthcare frontend modules from Angular to React, building reusable component libraries in Storybook aligned with Figma designs, cutting frontend development effort ~35%.
- Developed and maintained full-stack features using Node.js and C#/.NET REST APIs with SQL Server, enforcing input validation, secure auth flows, and API access controls in a regulated healthcare environment.
- Maintained ~80%+ code coverage across unit, integration, and E2E layers using Jest, React Testing Library, Selenium, NUnit, and xUnit.
- Built CI/CD pipelines with Azure DevOps and Docker, automating deployments and reducing production incidents ~35% within an agile delivery workflow.
Projects
Developers waste time digging through internal docs. Most AI tools send data to third-party servers — not viable for private or sensitive documentation.
Self-hosted RAG pipeline: documents are chunked with overlapping windows, embedded via OpenAI text-embedding-3-small, and stored in ChromaDB. Queries run semantic search over stored chunks and pass the top results to GPT-4o-mini with strict prompt constraints to prevent hallucination and enforce source citations.
P95 query latency under 1 second. Full observability with custom Micrometer metrics — P50/P95/P99 latency, error rates, and throughput tracked via Prometheus and Grafana dashboards.
Students had no structured system to visualize and plan their degree path — relying on spreadsheets, manual tracking, and advisor emails to figure out what to take next.
Built the Angular + TypeScript frontend with a dynamic course planning UI and OAuth-based authentication via Keycloak, connecting distributed Spring Boot microservices with independent PostgreSQL databases through an API gateway. Services containerized with Docker and configured with Kubernetes manifests for AWS deployment.
k6 load tests sustained 1,000 concurrent users and 100,000+ requests at 245ms median response time with 95%+ success rate.
Most student library system projects stop at a basic CRUD API. This one needed to actually work - with real borrowing workflows, queue management, and a deployable setup anyone could use.
Spring Boot backend with role-based access control, session-based authentication, and transactional borrowing workflows including 14-day rentals, 2x extensions, and an auto-managed holds queue. Deployed to AWS ECS Fargate behind an ALB and Render via a single GitHub Actions pipeline. Structured JSON logging with MDC correlation IDs and Prometheus metrics for observability.
Live on two platforms. k6 load tests: 100% success rate, 75ms avg and 109ms P95 response across 660 requests. CI/CD pipeline fully automated - push to main deploys everywhere.
Technical Expertise
Education
Let's ship
something real.
Graduating May 2026, available now. Looking for full-stack, frontend, or backend SWE roles where the work actually matters.