What Youll Do:
Develop and maintain Natural Language Understanding (NLU) models integrated into key company products.
Design and iterate on prompt templates and strategies using techniques such as few-shot, zero-shot, and chain-of-thought prompting.
Apply best practices from prompt engineering, including answer engineering (e.g., extractors, verbalizers, formatting).
Evaluate model outputs using modern benchmarking techniques (e.g., self-consistency, prompt ensembles).
Collaborate with cross-functional teams (engineering, product, clinical) to tailor LLM-based systems to practical use cases.
Monitor and improve model performance post-deployment with a focus on stability, safety, and domain alignment.
Contribute to the internal prompt engineering knowledge base and stay up to date with the latest research and community best practices.
Requirements
BSc in Computer Science, Engineering, Mathematics, Physics, or a related quantitative field. MSc is an advantage.
3+ years of hands-on experience in NLP roles with demonstrated production deployment of ML models.
At least 1 year of experience with LLMs (e.g., OpenAI, HuggingFace, Cohere, Anthropic, or similar).
Strong programming skills in Python with practical use of ML and NLP frameworks (e.g., HuggingFace Transformers, spaCy, scikit-learn).
Familiarity with prompt engineering principles, including prompt tuning, template construction, and response evaluation.
Experience with model monitoring and optimization in real-world settings.
Understanding of model limitations, prompt sensitivity, and output consistency.
Advantages:
Knowledge of ML-Ops practices and model deployment pipelines.
Experience with LLM safety, bias mitigation, or prompt hardening strategies.
Exposure to multilingual or multimodal prompting.
Familiarity with compliance frameworks in healthcare or other regulated environments (e.g., HIPAA, SOC2).
