I’m a Computer Science graduate and software developer focused on building scalable backend and full-stack systems.

I enjoy working with distributed systems, asynchronous processing, real-time applications, and AI/LLM-powered workflows.

Currently exploring LLM systems, RAG, AI agents, and production AI applications.

Languages: TypeScript, Python, JavaScript, C++, Java

Backend: Node.js, Express.js, FastAPI, REST APIs, Prisma, Mongoose

Frontend: Next.js, React.js, Tailwind CSS, shadcn/ui

Databases & Storage: PostgreSQL, MongoDB, MySQL, SQLite, Redis, Qdrant, S3-compatible storage

Systems & Streaming: Apache Kafka, BullMQ, WebSockets, Server-Sent Events, HLS, FFmpeg

AI & LLM: RAG, Agentic tool-calling, Multi-provider LLM gateways (Ollama · Groq · HF · OpenAI), Embeddings & vector search, LLM evaluation harnesses

DevOps & Tools: Docker, Linux, Git, GitHub Actions

A multi-tenant bug tracker built around two enforced guarantees: cross-tenant access always returns 404 (never 403), and the in-app AI assistant can only cite ids a tool actually retrieved for that tenant — never a fabricated reference.

• FastAPI agent service with native tool-calling (find_similar_bugs, get_bug, find_projects)
• Qdrant vector search for semantic duplicate detection and grounded retrieval
• Provider-agnostic LLM gateway (Ollama / Groq / HF / OpenAI) with a hard paid-API gate
• Kafka-driven real-time notifications over WebSocket
• 104 unit tests + a 41-case graded eval suite with committed baselines

Stack: Next.js · Express · FastAPI · MongoDB · Kafka · Qdrant · WebSockets

A full-stack video learning platform built around an asynchronous video processing and adaptive streaming pipeline.

• Video transcoding with FFmpeg and multi-bitrate HLS
• Background processing using BullMQ
• S3-compatible object storage using RustFS
• Course, lesson, quiz, payment and coupon systems
• Idempotent payment workflows and automatic tax calculation
• Instructor analytics and administrative management

Stack: Next.js · TypeScript · Express · PostgreSQL · Prisma · Redis · BullMQ · FFmpeg · HLS · RustFS

A deployed AI customer-support agent with live, testable retrieval-augmented answers.

• Live demo deployment (Vercel + Render + Neon Postgres)
• ~20 automated tests (Vitest/Supertest)
• Policy-as-data design and bounded context windows for cost control

Stack: TypeScript · Node.js · PostgreSQL · Gemini API

Shipped: RAG, agentic tool-calling, multi-provider LLM gateways, grounding/citation enforcement, LLM evaluation harnesses.

Exploring next: multi-agent orchestration, MCP, fine-tuning.

• LinkedIn: linkedin.com/in/ananyaa-koundal
• Email: ananyaa.koundal@gmail.com
• LeetCode: leetcode.com/u/Anna-K

⭐️ "Continuous learner, builder of impactful projects, and a believer in clean, scalable code."