LLM Engineer

Role Overview

We aim to bring about a new paradigm in medical image diagnostics; providing intelligent, holistic, ethical, explainable and patient centric care. We are looking for innovative problem solvers. We want people who can empathize with the consumer, understand business problems and design and deliver intelligent products. People who are looking to extend artificial intelligence into unexplored areas. We want Data Scientists who can be key contributors in a fast-paced dynamic product development environment for building artificial intelligence products and services. Your primary focus will be in applying deep learning and artificial intelligence techniques, and building high quality prediction and recommendation systems integrated with numerous APIs and big data backend.

Responsibilities

• Selecting features, building and optimizing classifiers and recommendation engines using deep learning techniques
• Use techniques from artificial intelligence/deep learning to solve supervised and unsupervised learning problems.
• Create, train, test and deploy various neural networks to solve complex problems.
• Develop and implement solutions to fit business problems which may include applying algorithms from a standard statistical tool, deep learning or custom algorithm development.
• Fine-tune state-of-the-art LLMs for specific use cases in healthcare technologies.
• Understanding the problem and creating appropriate solutions in accordance with the project requirements leveraging LLMs and other NLP artefacts.
• Working with vector databases and integrating relevant medical information to enhance reporting.
• Using frameworks like LangChain to create smart reporting pipelines integrating relevant information across formats and sources and turning them into intelligent reports.
• Finetuning and training Visual LLMs in conformance with technical and business needs.
• Understanding the performance, latency and memory needs to perform suitable solutions such as knowledge distillation and quantization Understanding the requirements and designing solutions and architecture in accordance with them.
• Participate in code reviews, sprint planning, and Agile ceremonies to drive high-quality deliverables.
• Maintain proper documentation for code, configurations, and architectural decisions.

Requirements

• Excellent understanding and hands-on of Deep Learning techniques such as ANN, CNN, RNN, LSTM, Transformers, VAEs etc.
• Must have experience with Tensorflow or PyTorch framework in building, training, testing and deploying neural networks
• Knowledge of the common data science and deep learning libraries and toolkits such as Keras, Pandas, Scikit-learn, Numpy, Scipy, spacy, NLTK etc.
• Good, applied statistical skills, such as distributions, statistical testing, regression, etc. Knowledge of data, data augmentation, data curation and synthetic data generation.
• Experience working with LLMs.
• Experience and understanding of Clinical LLMs will be an advantage.
• Excellent understanding of vector databases and experience working with them.
• Knowledge and experience of frameworks like LangChain.
• In-depth understanding of training of LLMs and finetuning them or domain specific tasks.
• Excellent understanding of prompting techniques such as one shot, chain of thought, etc. In-depth knowledge of quantization and its trade-offs.
• Proper working knowledge and experience of finetuning techniques such as PEFT and LoRa.
• Exposure to Agile/Scrum methodologies and collaborative development practices.
• Excellent verbal and written communication skills Excellent analytical and problem-solving skills with good attitude and keen to adapt evolving technologies

Education: BE/B Tech

Experience: 2-5 years