Responsibilities
Take over the existing vertical video generation pipeline (script → voiceover → visuals → subtitles → editing → auto-posting → tracking links) for YouTube Shorts, TikTok and Instagram Reels, understand it from the docs and code, and bring it to stable operation.
Grow the pipeline: improve content quality and virality, scale up generation volume, reduce cost and turnaround time per video.
Keep track of new generation models and services (video, image, voice, LLM), test them quickly and replace outdated parts of the pipeline with better or cheaper ones.
Launch and develop the virtual AI creator direction: a consistent character, voice, style and content plan.
Design and improve agents: write and debug prompts and sub-agent roles, validation between phases, context and cost control.
Build a dashboard for running generation and tracking content analytics (views, clicks, conversions, cost per video).
Automate publishing: platform integrations, posting from mobile devices, scaling channels.
- Write clean, testable, documented code, including unit and integration tests.
- Requirements
- Solid Python, basic bash.
Hands-on experience with LLM APIs (Claude, OpenAI, etc.) and agentic pipelines: a real project, pet project or hackathon entry you can show and talk through.
Experience using Claude Code, Cursor or similar tools as a working environment, not just as autocomplete.
Understanding of prompt engineering basics: defining an agent's role, limiting context, getting stable structured output.
Understanding of Git, team development and Docker basics.
Ability to work with REST APIs and webhooks, basic understanding of databases (SQL).
Understanding of vertical content: games, creators, streams, TikToks, Shorts, how video discovery works, what shadowbans are, how mobile posting farms are set up. A big plus if this is your hobby: your own channel, editing, streaming, experiments with AI generation.
Ownership: the ability to take a task end to end and deliver a result without external supervision.
Initiative: your own proposals on improving the pipeline, new models and automation, and the readiness to implement them yourself.
Ability to figure out someone else's system from code and documentation, reconstruct the logic and ask the right questions.
Independence in technical decisions and openness in communication: flag problems immediately.
Product thinking: measuring results by content metrics (views, retention, clicks) and cost per video.
Ability to explain the system's capabilities and limitations to non-technical colleagues.
Experimental mindset: fast hypothesis testing and the willingness to drop what did not work.
Desire to grow with the project: from stabilising the current pipeline to becoming its technical owner.
- English at the level of reading technical documentation (Intermediate+).
- Nice to have
- Experience with video and image generation via API (fal.ai, Kling, Runway, Seedance, Nano Banana, etc.), TTS and speech recognition (Whisper and similar).
ffmpeg: assembly, normalisation, subtitles, overlays.
Experience building reliable calls to paid APIs: retries, idempotency, cost tracking.
n8n or other workflow tools.
Telegram bots, YouTube / TikTok / Instagram / Discord APIs.
Android / mobile device automation.
Supabase or PostgreSQL.
Experience with GCP, Docker, Kubernetes.
Observability and monitoring (Datadog, OpenTelemetry).
Your own channel, pet projects, open-source experience or hackathon participation.
