Applied ML Engineer
Sentient · Remote · Remote - APAC ( Singapore ) · SF Only · Remote - APAC ( Hong Kong ) · Remote - EMEA ( UK ) · Austin ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Fully remote
- Python + PyTorch
- ML Evaluation focus
- Production engineering
- Startup environment
We seek an Applied ML Engineer to build systems at the intersection of machine learning research and production engineering. You will reproduce research, design rigorous evaluations, and transform promising methods into usable product features.
The role involves working across model internals, inference infrastructure, and user interfaces. Your goal is to create reliable, measurable systems that provide trustworthy evidence rather than just reproducing academic results.
What you'll do
- Reproduce and evaluate research methods using open-weight and API models
- Design evaluation datasets, scoring methods, and experiment harnesses
- Build and extend evaluation infrastructure including runners and reporting
- Turn research workflows into product experiences with traces and comparisons
- Investigate verification methods under model modification and evasion
- Ship production-quality systems with APIs, testing, and documentation
- Write technical reports distinguishing evidence from interpretation
What you'll need
- Strong Python engineering skills with PyTorch and Hugging Face Transformers
- Deep understanding of ML evaluation metrics, calibration, and reproducibility
- Ability to implement methods from research papers from first principles
- Experience building production software with APIs, databases, and deployment
- Comfort working with open-weight models and LLM inference systems
- Ability to work across backend and frontend boundaries using React/TypeScript
- High agency and ownership in a fast-moving startup environment
Nice to have
- Experience with model provenance, watermarking, or red-teaming
- Knowledge of activation analysis, probing, or model internals
- Familiarity with DSPy, LiteLLM, Temporal, Ray, or vLLM
- Skills in Next.js, React, TypeScript, or data visualization
- Experience serving open-weight models on GPUs with latency tradeoffs
Stack
- Python
- PyTorch
- Hugging Face Transformers
- React
- TypeScript
- Next.js
- DSPy
- LiteLLM
- Temporal
- Ray
- vLLM
- PostgreSQL
How you'll work
- Fully remote position
- Fast-moving startup environment
- Cross-functional collaboration expected