Job Description
This AI research scientist job in Bangalore is with the Agents team at Google DeepMind India. You will research large language models and build more capable autonomous agents with reinforcement learning. The team also applies its work to Gemini.
At a glance
- Position: Research Scientist, Agents, DeepMind (Google job ID 82988974666064582)
- Where: Bengaluru, Karnataka, India
- Degree: PhD in Computer Science, AI or a related field, or equivalent practical experience
- Experience: 8 years with ML frameworks such as JAX, TensorFlow or PyTorch, and ML publications
- Apply: open until filled (re-checked 29 October 2026)
What the DeepMind agents research scientist does
- Design, build and evaluate models, agents and software prototypes of large foundation models
- Push reinforcement learning and ML optimisation methods to build autonomous agents
- Lead work on new algorithmic architectures for more capable agents
- Present research results clearly, inside Google and outside
- Start and maintain collaborations with other research labs
- Work with Responsible AI teams so the work is developed ethically
Why this role matters
Few labs in India do frontier research on large models. This team studies how agent behaviour emerges from reinforcement learning at scale, and its results feed into Gemini, which hundreds of millions of people use.
What Google DeepMind is looking for
- PhD in Computer Science, Artificial Intelligence or a related field, or equivalent practical experience
- 8 years of experience with machine learning frameworks such as JAX, TensorFlow or PyTorch
- Experience with advanced deep learning and reinforcement learning
- A publication record at ML venues such as NeurIPS, ICML, ICLR, KDD or AAAI
Nice to have
- Multimodal learning, large language models or assistive AI agents
- Prompt engineering, few-shot learning, post-training and evaluations
- Model training and deployment
- Strong Python
Pay and location
The posting does not state pay. It is a research scientist role in Bengaluru. Apply from the job page on Google Careers; you will need to sign in with a Google account to submit.
How to apply
Apply through the official Google DeepMind job posting. Google DeepMind’s posting does not give a closing date, so the role is open until filled; ResearchJobs.in will re-check this listing on 29 October 2026. Check the posting for location and work-authorisation details before you apply.
See all our artificial intelligence jobs, or browse more research jobs on ResearchJobs.in.
Hiring institution: Google DeepMind
Official advertisement: www.google.com
How to prepare for this application
- Prepare a 20-minute talk on your best RL or LLM paper: the question, the method, the ablations and what failed
- Revise policy-gradient methods, PPO, and RL from human or AI feedback, since post-training is central to agents
- Be ready to design an evaluation for an agent: tasks, success metrics and how you avoid benchmark leakage
- Expect coding in Python and JAX or PyTorch; practise writing a clean training loop from scratch
- Read recent DeepMind work on agents and Gemini so you can say where your research fits
About Google DeepMind
Google DeepMind is Google's AI research lab. It was formed in 2023 by merging DeepMind and the Google Brain team. It develops the Gemini models and has research teams in several countries, including India.