An agentic document intelligence and research platform built with LangGraph and LangChain, featuring multimodal processing, vector search, citation-aware AI chat, and a context-rich assistant for AI, solutions architecture, and full-stack development.

Load test targeting the retrieval endpoint (/query/stream) with 10 concurrent virtual users over 5 minutes. Validates response times and error rates under normal operating conditions.

1 / 2
| Metric | Avg | Med | P95 | P99 |
|---|---|---|---|---|
| http_req_duration | 3772.47 | 3701.31 | 4875.46 | 5698.71 |

Click image to expand · 1 / 7
Branch to feature
developmentFeature work & test
featureMerge to development
featureTest on development
developmentMerge to main
developmentDeploy to production
mainBranch to feature
developmentFeature work & test
featureMerge to development
featureTest on development
developmentMerge to main
developmentDeploy to production
mainBranch from development to feature → develop & test → merge to development → run test suite → merge to main → GitHub Actions runs tests → deploy to AWS (backend), Vercel (frontend), Cloudflare (CDN/DNS).
Integrating workers for processing jobs with Redis Queue as a Broker.
Build a separate service for managing file conversion and management, decoupling from b-rag
Set up log aggregation pipeline
Add latency and request count metrics
Create monitoring dashboards
Implement metrics collection in API routes
Study observability patterns and anti-patterns