SURAJ KIRAN AIRI

Problem: Turning raw screencast or facecam footage into a usable rough cut means manually scrubbing through hours of audio to cut stutters, dead air, and rambling takes — tedious work that delays the creative editing that actually matters, and most AI editing tools lock footage into a cloud pipeline or a proprietary subscription instead of a professional NLE workflow.

Result: Built a full desktop application that transcribes footage, uses an LLM to make AI-driven keep/cut decisions, and exports either directly or as a fully editable OpenTimelineIO timeline for DaVinci Resolve — with a custom LCS-based text-to-timestamp aligner and waveform-aware boundary snapping so cuts land on natural pauses rather than mid-word or mid-syllable.

Problem: Building a robust personal AI assistant requires integrating multiple components: LLM providers, communication channels, tool execution, and maintaining conversation context, all while ensuring security, extensibility, and performance.

Result: Developed a fully functional AI assistant with 8 built-in tools, supporting multiple LLM providers and Telegram integration, demonstrating expertise in async Python development, API design, and agent-based architectures.

Problem: LLM agents often receive full schemas for all available tools on every turn, which bloats context, increases token cost, and hurts tool-selection accuracy.

Result: Built and published a Python package on PyPI that introduces a 3-meta-tool execution layer, reducing tool-context overhead by exposing tools on demand and using Python function-call syntax with AST-based validation. Added robust validation/error formatting and 90 automated tests.