Research Intern, AI

About the Role

IBM Research India offers Summer Internships to bright students from leading technical universities in India. This program provides industry experience and exposure to world-class technologies, allowing students to showcase their talent and work on current projects with mentors. The goal is to offer industry experience and identify future employees.

We are looking for a talented and highly motivated engineer/scientist to contribute to our research and development efforts across multiple frontiers of AI.

Key Responsibilities

• Work on one or more research areas, including:
• Building and scaling large language models.
• Agentic systems that work with small language models.
• Auto evolution of agents, including testing and debugging of agents.
• Software stack for AI Accelerator cards.
• Intersection of data and AI.
• Programmable language models.
• Solve problems with the help of mentors.
• Read, understand, and critically analyze research papers.
• Implement ideas from research papers.
• Reproduce research results and design controlled experiments.

Requirements & Tech Stack

• Python
• Solid understanding of Machine Learning and Deep Learning
• Familiarity with Large Language Models, Generative AI, Transformers, and NLP
• Hands-on experience with VS Code, Jupyter Notebooks, Git
• Exposure to AI Accelerator software stack
• Proficiency in AI libraries/frameworks: PyTorch
• Exposure to AI Agents, tool use, function calling, reasoning models

Qualifications

• Bachelor's Degree (B.E. / B.Tech)
• Master's Degree (M.E. / M.Tech, including Dual Degree programs)
• Ph.D.
• Minimum 70% or 7.0 CGPA and above in the pursuing degree.

Internship Duration: May to August 2027, maximum 3 months.

Preferred Qualifications

• Familiarity with concepts on linear algebra, probability, statistics and fundamentals of LLM models.
• Exposure to Triton and Hugging Face.
• Familiarity of AI Accelerator architectures, NCCL and compilers.
• Exposure to popular RL libraries and hands-on experience with RL training.
• Containerization (Docker/Kubernetes/Openshift).

Desirable Skills/Experience

• Familiarity with techniques such as LoRA/PEFT, knowledge distillation, reinforcement learning, or RLHF/RLAIF.