Saans AI

🚨 Hiring: AI/NLP Intern – Clinical Note Generation & LLM Evaluation (Remote) 📅 Start: ASAP | 🕒 Duration: 3–6 months | 💼 Unpaid (Potential for paid extension) We're a healthcare AI startup building agentic tools for medical documentation, patient monitoring, and insurance workflows . Our stack combines LLMs, multimodal AI, and real-time speech processing , backed by talent from NVIDIA, JP Morgan Chase, EY, and UC Health. We’re looking for a sharp AI/NLP intern to help us evaluate and fine-tune language models for SOAP note generation . You’ll work on real-world clinical data, build benchmarks, and integrate models into live systems. Perfect for students or early-career researchers passionate about applied AI in healthcare. 🔍 What You'll Work On: Fine-tune & benchmark LLMs (LLaMA, Mistral, MedAlpaca) Evaluate models using BLEU, ROUGE, BERTScore, and factuality Build NER + entity linking pipelines Deploy APIs (FastAPI) and work with front-end engineers Run error analyses, iterate prompts, explore new ideas ✅ Must-Have Skills: Python + HuggingFace Transformers + basic ML/NLP Understanding of prompting, attention, tokenization Jupyter or Colab for experimentation 💡 Bonus: FastAPI, React, full-stack experience Knowledge of clinical NLP or biomedical informatics 🔬 Evaluation Task: We shortlist candidates via a hands-on task using the ACI-Bench dataset (doctor–patient conversations → SOAP notes). You'll benchmark open-source LLMs and submit a short report or notebook. Details sent after initial screening. But, proactive candidates can already: Understand the Data : https://github.com/wyim/aci-bench/blob/main/data/challenge_data/train.csv Analyze the format of the doctor-patient encounters and how they map to the structured clinical notes. Explore the narrative and SOAP-style elements in the provided notes. Select and Apply LLMs - Choose one or more open-source LLMs (e.g., LLaMA, Mistral, MedAlpaca) that can be used to generate notes from the conversations. Decide whether to use zero-shot prompting, few-shot learning, or fine-tuning approaches. Define Evaluation Metrics - focus on hallucination and faithfulness 🎁 Perks: Work on real AI-for-healthcare systems Mentorship from experts High-impact portfolio project Strong LoR + possible transition to paid contract 📩 Apply by sending your resume + GitHub (if any) + a short blurb on your NLP experience to [founder@saans.ai]. Let’s build the future of clinical care—together. 🚀 #Hiring #AIInternship #NLP #LLM #HealthcareAI #GenerativeAI #RemoteJobs #InternshipOpportunity #MedTech #MachineLearning #TechForGood #DigitalHealth #AIInHealthcare #InternshipOpportunity