At ELEVANA GLOBAL, I architect event-driven backend systems that connect third-party SaaS APIs across client projects. These systems reduced manual operations effort by 60–80%.
I deliver production-ready MVPs using Docker on Linux and design asynchronous data pipelines with MongoDB. I also implemented isolated error handling so an integration failure doesn't disrupt the rest of a system.
Through open-source contributions to Agenta-AI, LiteLLM, and TraceRoot, I shipped merged fixes for LLM observability and provider connection flows. I also built CodeArena, a sandboxed code execution and AI interview platform, and developed machine learning projects including Synthetic Oracle.
