Hi, I'm kuschzzp.
I'm a backend developer who likes building the invisible parts of software: services, APIs, data flows, queues, caches, deployments, and all the small decisions that make a system feel fast and reliable.
Right now I'm focusing more deeply on the AI application layer: LLM APIs, agent workflows, RAG systems, tool calling, vector search, and how to connect AI capabilities with real backend architecture.
I care about systems that are:
- reliable enough to trust
- simple enough to maintain
- observable enough to debug
- flexible enough to evolve
- smart enough to help real users
|
Designing practical AI features around prompts, tools, structured outputs, memory, and backend service integration. |
Exploring agent planning, tool calling, task orchestration, human-in-the-loop flows, and reliable execution boundaries. |
|
Working with embeddings, vector search, document pipelines, retrieval quality, and answer grounding. |
Connecting AI services with APIs, databases, queues, caches, auth, observability, and production deployment practices. |
|
AI / Agents / RAG |
AI Runtime / Vector Search |
|
Backend Core |
Data / Cache / Messaging |
|
DevOps / Reliability |
Frontend / Tooling Touchpoints |
|
AI features should solve real workflow problems, not just look impressive in a demo. |
Clear APIs, clean data models, predictable performance, and boring production behavior matter. |
Good systems leave room for new users, new features, new data, and new failure modes. |
- LLM-powered backend workflows
- Agent orchestration and tool calling
- RAG pipelines, embeddings, and vector databases
- AI coding assistants and developer productivity tools
- Model context protocol and AI tool ecosystems
- Observability, evaluation, and safety for AI applications
|
Primary Languages |
Current Signal |
Backend first. AI forward. Build things that are useful, observable, and calm in production.