AI /ML Engineer Intern

Role Overview As an AI/ML Engineer, you will be a core technical member responsible for designing, developing, and deploying the AI engine that powers dynamic symptom assessment, adaptive questioning, probabilistic reasoning, and conversational health features. You will work closely with doctors, product teams, and full-stack developers to create safe, reliable, and scalable AI solutions. This is a high-impact role in a fast-moving startup environment. You will help shape the core intelligence of a product that can positively impact millions of users. Key Responsibilities Design and implement probabilistic reasoning systems (e.g., Bayesian networks or hybrid models) for symptom-to-condition assessment and differential diagnosis. Build and fine-tune conversational AI using LLMs (e.g., Llama, Mistral, or GPT variants) with Retrieval-Augmented Generation (RAG) for medical knowledge retrieval. Develop dynamic questioning logic that selects optimal follow-up questions based on information gain and user context. Create pipelines for processing unstructured inputs: text symptoms, lab reports (PDF/image analysis via OCR \+ vision models), and medication data. Preprocess, curate, and augment medical datasets while ensuring privacy and bias mitigation (especially for Indian demographics). Deploy and monitor models in production using MLOps practices (Docker, Kubernetes, CI/CD). Collaborate with medical doctors to validate model outputs, reduce hallucinations, and implement safety guardrails. Optimize for performance, latency, and cost on cloud infrastructure (AWS/Azure/GCP). Contribute to responsible AI practices: transparency, explainability, fairness, and compliance with health data regulations (DPDP Act, etc.). Experiment with hybrid approaches (rule-based \+ ML \+ LLMs) to improve clinical safety and accuracy. Required Qualifications \& Skills Experience: Hands-on experience in AI/ML, with at least 1–2 years in healthcare, medical AI, or NLP applications (preferred). Proven track record of deploying production-grade ML models (not just notebooks). Technical Skills: Strong proficiency in Python and ML frameworks: PyTorch, TensorFlow, Hugging Face Transformers, LangChain/LlamaIndex. Experience with LLMs, prompt engineering, fine-tuning, and RAG systems. Solid understanding of probabilistic modeling, Bayesian inference, or decision support systems. NLP expertise: entity recognition, intent detection, dialogue management. Data handling: Pandas, NumPy, vector databases (Pinecone, Weaviate, FAISS). MLOps \& Deployment: Docker, Kubernetes, MLflow, cloud platforms (AWS SageMaker, Azure ML, or GCP Vertex AI). Bonus: Experience with medical ontologies (SNOMED, ICD), vision models for document analysis, or multimodal AI. Education: Bachelor’s or Master’s degree in Computer Science, Artificial Intelligence, Machine Learning, Data Science, or related field. PhD is a plus. Soft Skills: Ability to work in a multidisciplinary team (with doctors and non-technical stakeholders). Strong problem-solving, analytical thinking, and attention to detail — especially around safety and edge cases in healthcare. Passion for healthtech and building products that “do no harm.” Preferred Qualifications Prior experience in symptom checkers, clinical decision support, or medical NLP. Knowledge of Indian healthcare context, multilingual AI (Tamil/English), or regulatory aspects of SaMD (Software as a Medical Device). Publications, GitHub projects, or contributions to open-source medical AI tools. Job Types: Full-time, Fresher Pay: ₹7,000\.00 - ₹10,000\.00 per month Application Question(s): * How many years of hands-on experience do you have in AI/ML engineering? * Have you worked on any healthcare, medical, or clinical AI projects (e.g., symptom assessment, diagnostic support, lab report analysis, or conversational health tools)? If yes, briefly describe one project and your role. (Short text – 200–300 chars) * Describe a recent ML project you built and deployed to production. What challenges did you face? * Which of the following tools/frameworks have you used in production? (Multi-select) → PyTorch / TensorFlow, Hugging Face Transformers, LangChain / LlamaIndex, RAG pipelines, Vector databases (Pinecone, Weaviate, FAISS), Docker \+ Kubernetes / MLOps tools? * What is your experience with Large Language Models (LLMs)? Have you done prompt engineering, fine-tuning, or built RAG systems? Give a short example. ( * How would you approach building a dynamic questioning system for a symptom checker (e.g., choosing the next best question based on user answers)? (Short text or optional) * In healthcare AI applications like symptom checkers or clinical decision support, what are the biggest risks (e.g., hallucinations, bias, safety)? How would you mitigate them? (Short text) * Have you worked with medical data, ontologies (SNOMED CT, ICD), or collaborated with doctors/clinicians? If yes, describe the collaboration. (Short text) * Why are explainability and safety guardrails especially important when building AI for health symptom assessment? (Short text) Education: * Bachelor's (Preferred) Work Location: In person