Elith Inc. is recruiting an engineering intern to help develop "GENFLUX," our generative AI quality evaluation platform.
Would you like to join us in building a next-generation AI foundation that systematically evaluates and controls the safety and quality of LLMs, centered on natural language processing (NLP) technology?π€
About Elith Inc.
Elith Inc. is a tech company focused on AI development centered on generative AI and LLMs.
With a track record of research acceptance at top international conferences and features in Forbes, we are rapidly expanding our business on the strength of "not stopping at research, but seeing it through to real-world deployment."
In particular,
- Accelerating enterprise adoption of generative AI
- Growth of the AI Safety and Responsible AI market
- Surging demand for quality assurance in enterprise settings
Against the backdrop of these trends, both our business and organization have entered an overwhelming growth phase.
Why "AI Safety" Now?
As the use of generative AI advances globally,
- Hallucination (lack of factual accuracy)
- Information leakage and prompt injection
- Bias and harmful expressions
- Regulatory compliance (EU AI Act, industry regulations)
risks such as these are becoming apparent.
Amid a trend of tightening regulation, starting with the EU AI Act,
demand is surging for "companies that can operate AI safely" rather than simply "companies that can build AI."
AI Safety is a field whose market value is set to grow significantly over the next few years.
Elith is seriously developing this field not as "defense," but as the foundational infrastructure that next-generation AI presupposes.
Product We're Building: GENFLUX
GENFLUX is a quality evaluation platform that checks the quality and behavior of generative AI responses in real time and automatically provides improvement suggestions.
- AI safety platform
- Guardrails β Prompt injection detection, input/output filtering, hallucination detection; Red Teaming β Exploring LLM vulnerabilities through adversarial prompt generation; compliance support (GxP / GDPR, etc.); toxicity and bias detection
- Quality Evaluation & Agent Layer
- RAG evaluation, secure RAG, various AI agent features
- No-Code Workflow Builder
- Build and run safe workflows using nothing but chat instructions
What Makes This Internship Special
β On the Front Lines of Implementing AI Safety Using NLP Technology
There are still not many companies in Japan seriously implementing AI Safety as a product.
In developing GENFLUX, you can work hands-on with NLP technologies such as the following:
- Detecting harmful expressions and evaluating bias through text classification
- Evaluating LLM output quality
- Detecting and defending against prompt injection attacks
- Designing evaluations for retrieval accuracy and faithfulness in RAG pipelines
In addition, at Elith you will encounter
- Enterprise projects
- AI Safety research at an international standard
- Connections to the physical AI domain
β themes that are rare even by global standards.
β‘ Experience FDE
Forward Deployed Engineer (FDE) β a new way of working for engineers, pioneered by Palantir and now fully adopted by top AI companies worldwide such as OpenAI and Anthropic. In 2025, FDE job postings recorded an 800% year-over-year increase, rapidly gaining attention as an engineering career path for the AI era.
An FDE is an engineer who dives deep into the customer's environment, directly connecting the product to the customer's problems.
Unlike traditional sales engineers or consultants, its defining feature is that you write production code yourself while solving the customer's business problems with technology.
At Elith, you'll have opportunities to deploy and customize GENFLUX while working directly with enterprise customers β even as an intern, you can practice this FDE-style way of working.
- Interview customers about their AI operations challenges and design technical solutions
- See the code you write run as an actual company's AI infrastructure
- Not just "build it and you're done" β experience development that commits to the customer's results
Become an engineer who can not only "write code" but also "solve the customer's problems." Environments where you can experience FDE-style work in Japan are still extremely limited.
β’ Overwhelming Speed of Growth
Because we're a small, elite team, the code you write is reflected in the product right away. You'll be assigned real tasks immediately after onboarding, and experience an early release after code review and testing.
β£ Working with Overseas Companies
Some tasks involve coordinating with overseas stakeholders, giving you an environment to challenge yourself with gathering information and communicating in English (support is provided based on your experience level).
Position
Engineering Intern
You will join the product development team and be directly involved in implementation.
Job Description
- Backend (FastAPI / Python) or frontend (Next.js / TypeScript) development
- Development of the AI Safety foundation (hallucination detection, harmful-expression classification, prompt injection countermeasures, compliance support)
- Development and improvement of quality evaluation features (RAG evaluation, etc.)
- Development of the no-code workflow builder
- Rapid development using tools such as Claude Code / Cursor
- Team development through code review and pull requests
Development Environment
- Language / Framework: FastAPI (Python), Next.js (TypeScript)
- Infrastructure: GCP, Docker, Terraform
- Tools: Git, Slack, AI tools such as Claude Code / Cursor
Requirements
Application Requirements
Required
- Basic knowledge of Python or TypeScript
- Development experience using GitHub (personal projects are fine)
- Able to commit at least 20 hours per week
Preferred
- Research or development experience in natural language processing (text classification, information extraction, NLI, etc.)
- Experience using NLP libraries such as Hugging Face Transformers / spaCy
- Knowledge of LLM evaluation methods (BLEU, ROUGE, BERTScore, etc.)
- Hands-on experience with prompt engineering
- Development experience with FastAPI / Next.js
- Experience with TDD (test-driven development)
- Development experience with RAG, LLMs, or generative AI
- Interest in AI Safety / Responsible AI
- Development experience on cloud platforms such as GCP / AWS
- Experience contributing to open source software
Working Conditions
- 20+ hours per week (flexible scheduling available)
- Fully remote OK
- Β₯1,300+ per hour (negotiable based on experience)
Selection Process
- Document screening
- Casual interview (online, approx. 30 minutes)
- Offer
How to Apply
Arm yourself with the latest tools and join us in building the next generation of AI infrastructure!
We look forward to applications from those who are serious about taking on the challenge of AI development.
βΆ Apply via our official website
You can apply directly through our company's form.
βΆ Apply via Wantedly
If you'd like to start with a casual conversation, click here.