Katy C.

Key Skills

Software

Google Cloud Vertex AI

Clickworker

Appen

Top Subject Matter

Sales voice training and real-time AI coaching.

AI-assisted video B-roll generation for product content workflows.

Agentic tool orchestration and supervised execution workflow design for AI systems.

Top Data Types

Audio

Video

Top Task Types

Transcription

Prompt + Response Writing (SFT)

Function Calling

Freelancer Overview

CCPractice / Resonatr — Real-time AI voice coaching platform (Personal Project, 2024–Present). Brings 3+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Deepgram, Claude API, and Google TTS. Education includes Associate of Applied Science, Maestro College. AI-training focus includes data types such as Audio, Computer Code, and Programming and labeling workflows including Transcription, Prompt + Response Writing (SFT), and Function Calling.

Labeling Experience

Agent Team Orchestration Platform — 15-agent AI platform with MCP-compatible architecture (Personal Project, 2025–Present)

Function Calling

Designed and implemented an MCP-compatible multi-agent orchestration platform that routes agent tools and governs execution via permission gates and approval workflows. Used shared memory and persistent state to maintain context across agent runs, enabling consistent behavior in iterative task execution. Built real-time dashboards and event streaming to monitor and validate agent performance during orchestration. • Routed tasks across 15 specialized agents using an MCP-style tool registry. • Implemented approval workflows and permission gates for controlled execution. • Added persistent memory to avoid duplicate work across sessions. • Built WebSocket-based real-time monitoring and per-agent status reporting.

2025 - Present

RawBlockAI — Multi-agent B-roll video generation pipeline (Personal Project, 2024–Present)

Prompt + Response Writing (SFT)

Built multi-agent systems that research product information and generate targeted instructions/prompts to drive downstream video generation. Designed an agent pipeline that coordinates task research, prompt creation, and production of B-roll outputs as an iterative generation loop. Emphasized orchestration quality and repeatable outputs using containerization for production runs. • Created a 5-agent pipeline for research, prompt targeting, and B-roll generation. • Coordinated agent outputs to form effective prompt inputs for the video model. • Containerized the pipeline with Docker for consistent production execution. • Leveraged multi-agent orchestration patterns to manage generation tasks end-to-end.

2024 - Present

CCPractice / Resonatr — Real-time AI voice coaching platform (Personal Project, 2024–Present)

AudioTranscription

Developed an end-to-end real-time voice AI coaching workflow that transcribes user speech with low latency and feeds it into live evaluation. Used continuous audio streaming to convert spoken input into text for subsequent AI scoring and coaching responses. Focused on correctness and responsiveness for real-time training feedback during interactions. • Implemented streaming audio pipeline with sub-second latency. • Used Deepgram for real-time speech-to-text and Claude API for live evaluation. • Delivered coaching feedback during conversations as part of the training loop. • Built supporting app UI/visualization and authenticated user flows via Supabase.

2024 - Present

Education

Work History