Available for work
Asim
Delvi
Software Engineer & AI Builder
Building AI-powered web apps that scale. Automating complex workflows with Python, FastAPI, Next.js, Docker and Kubernetes — turning ideas into impactful software.
~/asim — zsh
· Asim Delvi — Portfolio Terminal v1.0 ·
──────────────────────────────────────────────
Type help to see available commands.
Use ↑ ↓ for command history, Tab to complete.
❯
scroll
Spotlight
★ Featured Project
VidWhiz
AI platform with a RAG pipeline (GPT + Pinecone Vector DB) for YouTube video Q&A and summaries, plus a Chrome extension for in-page interaction.
NestJSNext.jsOpenAI APIPinecone Vector DBTypeScriptChrome Extension
Tech Stack
Languages
JavaScriptTypeScriptPython
Frameworks & Libraries
ReactFastAPISQLAlchemyZustandTailwind CSSAnt DesignNext.jsLangChainLangGraphExpress.jsPrismaRedux
Tools & Infrastructure
DockerKubernetesAWS EKS / FargateKustomizeAWS S3 / LambdaAWS MediaConvertGitLinux
Databases
PostgreSQLMongoDBRedis
Experience
Software Engineer
KrtrimaIQ
- Led the engineering track for an AI-powered SaaS hiring platform with automated JD generation, resume shortlisting, and AI-driven candidate interviews using React, Zustand, Ant Design, FastAPI, and PostgreSQL.
- Architected a multi-org hierarchy with client-vendor relationships and 5-role RBAC using Clerk, designing PostgreSQL/MongoDB schemas with complex table relationships and foreign keys.
- Designed and implemented 3 end-to-end insurance denial resolution workflows with 10 complex business rules per workflow — parsing CPT/ICD codes, correcting them, and resubmitting claims, reducing manual intervention by 70%.
- Built a secure video streaming pipeline using AWS S3, Elemental MediaConvert, and pre-signed URLs, reducing initial video load latency by 40% through chunked delivery.
- Set up and managed an AWS EKS Fargate cluster serving 8–10 services with Kustomize manifests, cron jobs, ingress controllers, namespace isolation, RBAC, and secrets management.
- Developed a RAG-based application using LangChain that extracts content from PDFs and videos to auto-generate structured quiz questions in JSON format.
- Replaced sequential database queries with MongoDB aggregation pipelines for complex multi-document lookups, reducing round trips and improving data retrieval performance.
- Implemented JWT-based role authentication and Microsoft SSO (MSAL) for multi-tenant platforms.
ReactZustandAnt DesignFastAPIPostgreSQLMongoDBLangChainAWS S3MediaConvertEKS / FargateKustomizeClerkJWTMSAL
Full Stack Developer
Freelancing
- Rebuilt a visiting card e-commerce platform from Angular to React, delivering an admin dashboard, blog CMS with rich text editor, order management, and Stripe/Razorpay billing with geo-based currency detection.
- Developed a canvas-based card customization editor with drag-and-drop positioning, real-time preview, multi-layer text styling, image uploads, and IndexedDB storage for large assets.
- Built a document management system with Google Drive integration, file/folder CRUD, trash/restore, move operations, user search, infinite scroll pagination, and multi-file upload with progress tracking.
- Implemented SEO optimizations (lazy loading, structured data, canonicalization) achieving 90+ Lighthouse scores across mobile, tablet, and desktop breakpoints.
ReactNext.jsStripeRazorpayCanvas APIIndexedDBGoogle Drive APITailwind CSSRedux
Projects
01
VidWhiz
AI platform with a RAG pipeline (GPT + Pinecone Vector DB) for YouTube video Q&A and summaries, plus a Chrome extension for in-page interaction.
NestJSNext.jsOpenAI APIPinecone Vector DBTypeScriptChrome Extension
03
Protobuf Benchmark Service
Go web service benchmarking JSON vs Protobuf serialization, demonstrating significant payload size and parsing speed improvements. Containerized with Docker Compose.
GoGo FiberMongoDBProtobufDocker Compose