L
LILT AI builds translation and data labeling infrastructure for enterprises and governments needing multilingual AI systems. The stack reveals heavy ML focus—PyTorch, Transformers, LangChain, LlamaIndex, RAG—paired with production ops tools (Datadog, Wiz, GCP/AWS). Hiring is heavily weighted toward data roles (24 open positions) and specialized evaluation workflows, suggesting the core problem isn't translation volume but data quality: reducing hallucinations, detecting cultural drift, and validating model safety across languages.
Tech Stack 38 technologies
Core StackDatadog BigQuery Salesforce AWS Python GitHub RAG Ashby PyTorch LangChain Rippling Adobe InDesign Adobe Illustrator Adobe Photoshop Drata Vanta Wiz GCP Bash Git LinkedIn Tipalti Payoneer Transformers LlamaIndex DATEV Carta FrameMaker QuarkXPress Microsoft Office+8 more
What LILT AI Is Building
◆Challenges
- Cultural mismatch in ai outputs
- Style drift
- Reducing hallucinations in ai outputs
- Ensuring safety
- Improving ai model quality
- Clinical accuracy in translated content
- Surface rubric gaps
- Detecting cultural mismatch
- High-quality ai evaluation
- Improving model deployment readiness
▲Active Projects
- Ai voice model training
- Multilingual ai evaluation workflows
- Model monitoring and regression testing
- Red-teaming and safety evaluation
- Red-teaming and security review
- Text review & correction
- Video content understanding
- Prompt evaluation
- Reviewing translated dental content
- Calibration and qa processes
Hiring Activity
120 roles · 120 in 30d
Department
Seniority
Notable leadership hires: Head of Security
Company intelligence
