- 🔭 I’m currently working on Agentic AI, LLMs, and Orchestration Pipelines
- 👯 I’m looking to collaborate on Cloud-Native Architecture, AGI, and Web3
- 👨💻 All of my projects are available at avinashwebfolio.netlify.app
- 📫 How to reach me: m.avinashkumar96@gmail.com
Ordered by specialization and architectural depth:
- Agentic AI & Processing: Python, OpenAI, LlamaIndex, Meta/Facebook APIs (WhatsApp Cloud API), AWS Textract, Azure AI
- Backend, Systems & Serverless: Golang, Node.js, Serverless Architecture, gRPC, Protocol Buffers (Protos), FastAPI
- Frontend Frameworks: Next.js, React
- Cloud & Orchestration: AWS, GCP, Azure, Docker, Kubernetes, Fly.io, QStash (Upstash)
- Databases & Search: Vector DBs, Elasticsearch (ELK), PostgreSQL, MongoDB
Note
Privacy & Contributions Notice: Professional projects listed here were developed using secure enterprise accounts provided by my employers to comply with privacy policies. As this is a personal GitHub account, the contribution graph only reflects occasional open-source work and personal experiments, not my daily commit history or full production output.
An automated agentic workflow designed for extracting unstructured invoice documents and structure-loading them into data stores.
- The Engine: Powered by Python and AWS Textract for OCR, combined with OpenAI for semantic text-to-JSON schema parsing.
- The Pipeline: Leverages a high-performance Go gRPC Protos pipeline to stream and validate transaction payloads.
- Orchestration: Integrated with QStash queue mechanisms to ensure robust, decoupled, and asynchronous task retry handling.
An enterprise-grade omnichannel conversational assistant designed to synchronize physical store inventories with AI vector systems.
- Core Systems: Built on FastAPI (Python) and integrated seamlessly with Chatwoot to manage live agent/AI handoffs over WhatsApp.
- Vectorization Engine: Periodically ingests product listings directly from legacy POS systems, generating semantic embeddings stored and indexed inside ELK (Elasticsearch) for instant, natural language product matchmaking.
- AI Backend: Powered by Azure AI for predictable, low-latency contextual reasoning.
A developer-focused DevOps platform and CLI designed to strip away cloud deployment complexities.
- Architecture: Written entirely in Golang utilizing GraphQL APIs for deterministic configuration streaming.
- Cross-Cloud Engine: Abstracts container orchestration across Kubernetes, deploying containerized application states dynamically over infrastructure providers like Fly.io, AWS, GCP, and Azure.
