Youll join a small, highly autonomous team with the agility, ownership, and pace of a startup, working alongside world-class engineers and researchers to build the next generation of conversational AI. Youll have access to large-scale real-world data, the compute and resources to pursue groundbreaking research, and the freedom to move fast while helping shape Harmonys AI vision.
What youll do
Research novel algorithms and design new deep learning architectures to advance our in-house NLP and conversational models.
Develop methods for translating customer goals into structured representations with controlled states and real-time observations.
Help design the reasoning layer that orchestrates our AI models efficiently alongside a range of LLMs, enabling each agent to adapt its policy throughout a conversation while remaining aligned with customer objectives.
Build analytical tools that trace our AI decision process across a vast corpus of conversations, providing clear visibility into their reasoning and behavior.
Collaborate closely with engineers and product teams to translate cutting-edge research into production at scale.
Proactively lead new ideas and technical initiatives.
MSc or PhD in Computer Science, Mathematics, Physics, Electrical Engineering, or a closely related field.
Proficiency in Python.
Strong theoretical background in machine learning and deep learning principles.
Solid understanding of NLP concepts and relevant linguistic principles (co-reference, natural language inference, grammatical relations, and similar).
Experience using statistics and information theory to evaluate AI models and conduct EDA.
4+ years of experience conducting NLP research with deep learning models and frameworks (PyTorch, TensorFlow, and similar).
Experience shipping AI systems from research to production.
Experience with conversational AI agents is a big plus.
Publications at top-tier conferences (ACL, EMNLP, NeurIPS, ICML, CVPR) are a plus.
