Kirillitsa, a localization studio, has been voicing over 1+ million minutes of content for Disney, Netflix, Sony, and others since 2009.
We are transforming a traditional studio into an AI-first company, and development is conducted with AI agents using our proprietary methodology (GRACE Framework).
Product to Work On
VoxQube Actor Portal, a secure remote actor recording SaaS. Professional actors gain access to protected video sessions (DRM, watermark, signed links), read text synchronized with video, and record voiceovers directly in the browser (studio quality, real-time audio streaming to the server). Under the hood: Next.js fullstack, PostgreSQL, AWS (EC2, S3, CloudFront, Batch), Terraform, ffmpeg video and audio processing pipeline.
The platform is undergoing certification for Trusted Partner Network (TPN), an industry standard for content protection, mandatory for working with major studios. This is an active part of the current work, not a one-time formality.
What You Will Do
- Develop a fullstack system on Next.js (App Router, Server Actions, route handlers)
- Build a backend with a role-based access control (RBAC) model and circuit isolation
- Work with secure video streaming and in-browser audio recording (Web Audio, WebSocket, ffmpeg)
- Conduct security hardening as part of preparation for TPN certification: DRM, watermarking, audit logging, addressing pentest findings, and activating committed but not yet active protective mechanisms in production
- Integrate AI agents into the development process (prompting, orchestration, skills)
- Configure and maintain infrastructure and CI/CD that are actually used, not just lying in the repository
Requirements:
Backend:
- PostgreSQL (schemas, indexes, migrations, transactions)
- Next.js as a backend layer (route handlers, middleware, Server Actions)
- JWT + role-based access (RBAC)
Frontend / Fullstack:
- Proficient TypeScript
- Next.js (App Router, middleware, Server Actions)
- Zod (shared DTO)
Infrastructure / DevOps:
- Solid understanding of AWS infrastructure (EC2, S3, RDS, CloudFront, Secrets Manager)
- Terraform configuration
- S3 (presigned URLs, multi-bucket)
- Docker
- CI & CD: experience in setting up and maintaining build and delivery pipelines (Jenkins, GitLab CI/CD or similar)
Additional:
- Basic knowledge of Python
AI:
- Understanding of semantic code / knowledge graph
- Practical experience with vibe-coding. Not just "tried Copilot", but regular work with AI agents in autonomous multi-step tasks (from defining a development phase to code review of the result), not just point-in-time autocompletions
- Ability to write skills for agents
- Prompt engineering and agent autonomy management (when to give freedom, when strict boundaries are needed)
- Experience with AI tools like Claude Code / Cursor, specifically in projects, not in test tasks
Will be a plus:
- Webhooks (HMAC, IP validation)
- Multi-tenant SaaS (companyId isolation)
- React Query / SWR
- Radix UI / headless UI
- Logging (Pino / Winston) + Sentry
- Web Audio (AudioWorklet, WebSocket streaming) / Canvas
- Auth (TOTP, Magic Link)
- Working with DRM and protected video streaming (CloudFront Signed URLs, CENC)
How We Work
- Given a task → you choose the solution yourself
- We test quickly and move forward
- AI is part of development, not just for playing around
What You Will Get
- A clear system: architecture, knowledge graph, module contracts, verification plan
- AI methodology (GRACE Framework)
- Onboarding from a developer (no chaos)
- Real growth in AI and architecture
Who This Is Not For
- Those who need constant supervision
- Those with no independent development experience
- Those who think "AI is interesting, but haven't tried it"
How to Apply
Write briefly:
- what you have actually done
- how you use AI
- a link to your projects
