Applied AI Engineer LLMs & Vision

About the Role

We are building AI systems that turn medical imaging into draft radiology reports, combining frontier and locally hosted models with evaluation and validation workflows before a radiologist signs off.

As an Applied AI Engineer, you will work across the stack — from LLM generation and evaluation to retrieval, multimodal models, and personalization. This is a hands-on role for someone who enjoys building, experimenting, measuring results, and working directly with data.

What You'll Work On

·      Build and improve LLM-based systems for generating and structuring long-form clinical text across imaging modalities including X-ray, CT, MR, ultrasound, and mammography.

·      Design prompts, evaluation frameworks, golden datasets, automated graders, including LLM-as-a-judge, and regression gates across thousands of de-identified reports.

·      Experiment with multimodal and on-premise vision-language models, benchmark competing models, and determine which model is best suited for each task.

·      Build retrieval-augmented generation (RAG) systems end-to-end, including chunking, embeddings, indexing, ranking, and retrieval-quality evaluation.

·      Develop personalization systems using preference memory and feedback loops to adapt model outputs to individual users.

·      Work closely with the engineering and clinical teams to understand failure modes, investigate data, and continuously improve model performance.

What We're Looking For

·      Strong Python and computer science fundamentals.

·      Pursuing or recently completed a B.Tech/M.Tech, with a strong preference for candidates from IITs, NITs, or BITS.

·      Final-year students and recent graduates are welcome to apply.

·      Hands-on experience building with LLMs, such as prompts, RAG, agents, model APIs, or similar applications.

·      Ability to design meaningful evaluations and make decisions based on data and measurable results, not intuition alone.

·      Strong analytical thinking, curiosity, ownership, and a willingness to dig into data and understand why a system succeeds or fails.

Relevant Experience That Stands Out

Experience in any of the following areas is valuable:

·      Medical imaging AI: detection, classification, segmentation, or report generation involving X-ray, CT, MR, ultrasound, or mammography.

·      Medical imaging research, internships, theses, competitions, or independent projects.

·      DICOM, PACS, or public medical-imaging datasets and challenges.

·      Vision-language models, computer vision, or multimodal AI.

·      LLM evaluation, observability, and benchmarking.

·      Vector databases, embeddings, and semantic search.

·      Modern deep-learning frameworks.

·      Interest in clinical data, healthcare workflows, and applied AI.

Why This Role

You'll get the opportunity to work on a real-world AI problem at the intersection of LLMs, computer vision, healthcare, and clinical workflows — while working across experimentation, evaluation, retrieval, and production-oriented AI systems.

This role is best suited to someone who wants to go beyond simply using AI APIs and actually understand, evaluate, and improve how AI systems behave.