FullStack AI Developer
Full-time FullStack AI Developer position to build next-generation voice AI infrastructure
About VoicePing
VoicePing is revolutionizing global business communication by developing the foundational Voice AI Model through cutting-edge technologies. Founded in 2019, we’ve grown to serve 1000+ enterprise customers with our industry-leading speech translation infrastructure.
Learn More:
- Our Product
- Product Development Journey
- Insights from 500 Global’s J-StarX Program
- Alchemist Accelerator Participation
- Company Profile
Company Achievements
- 1,000+ paying enterprise customers
- $2.3M USD total investment as SEED
- Recognition from global accelerators (500 Global, Rainmaking APAC, AlchemistX)
Technical Infrastructure Overview
Core AI Stack
| Component | Technology |
|---|---|
| Speech Recognition | Custom-trained transformers optimized for Asian languages |
| Translation Engine | Fine-tuned LLMs with domain-specific training |
| Text-to-Speech | Natural emotional synthesized audio generation model |
| Language Models | Specialized models for business terminology and technical vocabulary |
| Audio Processing | Real-time audio streaming with custom DSP pipelines |
Development Stack
Frontend:
- Core: React, TypeScript, Flutter (iOS, Android)
- Real-time: WebSocket, WebRTC for audio streaming
- UI Components: Ant Design, Flutter
Backend:
- Primary: Node.js (TypeScript), Python (FastAPI), Golang
- APIs: REST
DevOps & Infrastructure:
- Cloud Platforms: AWS, GCP, Azure, VPS, OnPremise
- CI/CD: GitHub Actions, Azure Pipelines
AI/ML Pipeline:
- Tools: PyTorch, Hugging Face, Llama, Unsloth, Qwen
Core Responsibilities
Infrastructure Development
- Build and maintain real-time audio processing pipelines
- Implement high-performance API endpoints for translation services
- Develop and optimize WebSocket connections for live translation
- Create and maintain containerized microservices
AI Integration
- Collaborate with ML engineers on model deployment
- Implement model serving infrastructure
- Optimize inference pipelines for low latency
- Develop monitoring systems for model performance
AI Model Creation
- Design data pipeline strategy and collect high-quality data
- Adjust AI model parameters (hyperparameters, architecture) to improve VRAM efficiency and inference speed
- Create model evaluation strategy, analyze results, and determine the right direction
- Research latest OSS models and papers to improve current models
- Publish research articles and contribute to OSS ecosystem
Application Development
- Build responsive web interfaces for real-time translation
- Implement offline translation capabilities
- Create developer tools and SDKs
- Develop internal dashboards for system monitoring
Technical Requirements
Must Have
Professional experience is not required. We welcome applicants with strong web application coding skills, demonstrated through personal, study, open-source, or work projects.
- Ability to build and debug web applications with React/TypeScript or a comparable stack, and explain your own code
- Understanding of frontend state, APIs, backend logic, and data storage; Node.js, Python, or similar technologies are welcome
- Ability to use Git and tests to investigate problems and verify your changes
- Willingness to learn our cloud, deployment, and AI tools
Nice to Have
- Audio and Video Digital processing
- Knowledge of ML model deployment and inference optimization
- See our research: Translation Inference Throughput , Translation Bottleneck
- Background in NLP or speech recognition or translation research
- Understanding of high-performance computing and scalable data processing
- See our research: Go WebSocket Proxy
Current Technical Challenges
- Optimizing real-time audio processing for low latency
- E2E model creation for emotional speech translation
- E2E model creation for AI voice agent bot
- Scaling ML model inference to reduce latency and energy use
- Implementing efficient offline translation
- Building robust monitoring systems to support our infrastructures
- Developing SDK for third-party integrations
- Large-scale multilingual training data collection and creation
Growth Opportunities
- Deep dive into ML infrastructure and ecosystem
- Learn advanced audio processing
- Machine learning training framework
- Data pipeline for AI model training data
- Master distributed systems
Benefits
- Health insurance
- Regular technical workshops
- Work directly with the founder and lead engineer
Robotics: an additional engineering role
We are also recruiting a dedicated Robotics & Physical AI Engineer . That role covers mechanics, electronics, embedded control, and software integration. The full-stack role above continues to focus on our voice AI products and infrastructure.
Five values that shape how we work
These values guide our engineering, enterprise sales, and the way we build a team.
1. Blue Ocean Through the Third Door
Find unconventional paths to new markets. Look beyond crowded feature competition and create new value for customers. Aim beyond survival, toward meaningful work and self-actualization.
2. Clock-Builders
Build an enduring organization as well as products. Invest in data, processes, and people whose value grows over time. Develop teammates who go on to lead the industry.
3. Whole Product
Own the customer’s complete job. Bring the product, integration, and support together, with the accuracy, speed, and flexibility the task needs.
4. Speed + Learning
Ship fast and learn faster from customers. Sweat the details that matter and make deliberate tradeoffs elsewhere.
5. Selected DNA
Bring curiosity, resilience, and personal drive. Share the mission and take ownership. Working side by side, learning together, and sharing a table build the conviction that sustains the team.
Work Environment and Compensation
- Employment: Full-time
- Workplace: Yokohama–Shinagawa area · Fully on-site
- Annual salary: JPY 5.5–10 million, discussed based on your current salary.
How to Apply
Send your application to Akinori Nakajima by LinkedIn direct message, including the details below.
- The role you are applying for and a brief introduction.
- Your skills and relevant work, study, or personal projects, with a CV, portfolio, or project links.
- The value you connect with most and an example of how you put it into practice.
Open Nakajima’s LinkedIn profile
A free LinkedIn account is sufficient. If you are not connected, send a connection request first. Once it is accepted, send your application by message.
If LinkedIn is not working for you, email your application to: nakajima@voice-ping.com