Daria
Zhmud

AI animation and video production studio building branded content, short-form video, AI avatars, commercials and creative automation workflows.

• Lead the creative team and organize the studio process: roles, briefs, creative direction, production tasks, review cycles, approvals and final delivery.
• Designed and implemented an AI-first production workflow covering idea intake, scripting, visual direction, asset generation, editing handoff, quality control and client delivery.
• Built repeatable production pipelines for Reels, AI commercials, avatar videos and branded assets, turning one-off AI experiments into structured studio workflows.
• Built AI-assisted systems for analyzing creative performance and improving KPI: content hypotheses, production feedback, performance signals, iteration loops and decision support for the team.
• Integrated LLM workflows, Telegram bots, Notion databases, parsers and content pipelines into a unified production infrastructure.
• Implemented human-in-the-loop processes where AI accelerates production while creative review, brand checks and approvals stay controlled.

Telegram-based content assistant for DariDora content operations.

• Routes requests across Telegram, Instagram, Reels and Threads instead of treating every format as generic content.
• Generates drafts, hooks and ideas, saves structured content into Notion and supports review, revision and publishing preparation.
• Uses editorial feedback, brand style and previous corrections as learning signals for future drafts.

Internal system for running AI studio work through specialized roles, lead workflows and approval logic.

• Architected specialized AI-agent roles for lead research, account analysis, outreach drafting, approval flows and safe client communication.
• Designed lead parsing workflows for data collection, classification, account enrichment, prioritization and outreach draft preparation.
• Created a board-level decision layer for pricing, products, leads, conflicts and strategic choices.

Project-specific platform architecture for producing content with an AI model before the content pipeline starts.

• Packaged the project into a character production OS: identity canon, visual references, face/hair/outfit/location anchors and reusable scene presets.
• Designed an agent workflow that turns a builder payload into prompts, negative prompts, reference selection, ComfyUI runs, QA checks, retry logic and saved recipe reports.
• Added production accounting: approve/reject statuses, generation history, cost ledger and effective cost per approved asset.

Agent-assisted infrastructure for device routing and account warmup before scaling content distribution.

• Built a standalone TikMatrix / IgMatrix layer for phone-farm health checks, device/account channels, warmup runs and task-status monitoring.
• Separated warmup-only mode from publishing so accounts can be prepared safely before any posting automation is enabled.
• Configured conservative engagement presets for cold accounts: search-based warmup, likes/follows/collects limits and no comments by default.

Personal brand around practical AI, agent workflows, content systems and the real work of building an AI-native studio.

• Built a multi-platform publishing workflow across Instagram, Reels, Threads, Telegram and Notion.
• Designed a Reels-led growth system with hooks, scripts, lead magnets, code-word CTAs and DM automation.
• Turns technical AI processes into accessible content for founders, creators, expert businesses and teams.