Large Model Application Development Engineer

Job Responsibilities

1. Responsible for the R&D and Application of Large Models in the Mental and Psychological Vertical Domain: Deeply involved in algorithm development for core scenarios such as AI pre-diagnosis and CBT (Cognitive Behavioral Therapy), mining and utilizing professional medical data (such as de-identified conversations, scales, and medical records), and constructing industry knowledge graphs.

2. Build Professional AI Pre-Diagnosis Models: Responsible for algorithm research on core NLP tasks such as multi-turn dialogue, intent recognition, symptom mining, and emotion perception. Combining cutting-edge large model training and optimization methods (such as SFT, GRPO) to improve the professionalism, accuracy, and empathy of the model in pre-diagnosis scenarios.

3. Develop CBT Agent Architecture: Responsible for designing and implementing AI Agents capable of performing structured CBT interventions. Overcoming technical challenges such as multi-step reasoning, dynamic memory, and RAG (Retrieval-Augmented Generation) to enable the Agent to understand and guide users through the CBT process (such as cognitive restructuring and thought journaling).

4. Ensure Model Safety and Value Alignment: Researching and implementing AI safety and alignment technologies, focusing on the specific characteristics of the mental and psychological field (such as crisis intervention, user privacy, and ethical red lines), conducting specialized safety reinforcement and value alignment of the model to ensure the reliability and beneficence of AI applications.

5. Build Evaluation and Iteration System: Participate in building and improving AI evaluation benchmarks in the mental and psychological field, evaluating and iterating on model and Agent performance from multiple dimensions such as professionalism, empathy, safety, and task completion.

Job Requirements

1. Bachelor’s degree or above in Computer Science, Artificial Intelligence, Biomedical Engineering, or a related field.

2. Solid foundation in machine learning/deep learning, proficiency in Transformer architecture, and familiarity with the principles and technologies of mainstream large models (LLM, VLM).

3. Proficient in Python, proficient in at least one deep learning framework, and possesses strong engineering implementation, algorithm debugging, and performance optimization skills.

4. Experience in NLP or vertical industry large model Pre-train, Finetune, Inference, and other optimizations.

5. Excellent learning ability and English literature reading ability, able to quickly follow up and apply cutting-edge technologies.

6. Strong interest in mental and psychological health, CBT, and other fields, eager to solve real-world problems, pursue the user value of AI technology, and possess good empathy and humanistic care.

7. Good teamwork and communication skills, proactive, and has strong analytical and problem-solving abilities.

Preferred Qualifications

1. Experience in related industries or projects such as medical health, online consultation, and psychological counseling is preferred.

2. In-depth research or project experience in NLP, dialogue systems, affective computing, Agent, or knowledge graph fields is preferred.

3. In-depth experience in long-term memory, RAG, reinforcement learning (especially RLHF) is preferred.

4. Leading impactful AI projects or publishing papers in top AI conferences (such as ACL, NeurIPS, ICML, etc.) is preferred.