Research Engineer - Multimodal Companion Agent at DeepMind

Tokyo, Japan

DeepMind Logo
Not SpecifiedCompensation
Senior (5 to 8 years)Experience Level
Full TimeJob Type
UnknownVisa
Artificial Intelligence, TechnologyIndustries

Requirements

  • Expertise in LLM post-training and evaluation
  • Experience with multimodal domains (vision, audio, text)
  • Proficiency in developing and optimizing multimodal AI models
  • Skills in building and maintaining robust data pipelines for training and evaluation
  • Ability to design, implement, and run experiments using metrics, prompt engineering, and few-shot learning
  • Knowledge of implementing algorithms for analyzing user interactions via vision and audio
  • Experience working closely with research scientists and engineers in a collaborative environment
  • Passion for AI, human-computer interaction, and agentic technologies

Responsibilities

  • Translate research concepts into practical implementations by developing and optimizing multimodal AI models, and building and maintaining robust data pipelines for training and evaluation
  • Design, implement, and run experiments to evaluate the performance and robustness of multimodal companion AI agents, using metrics and techniques like prompt engineering and few-shot learning
  • Implement algorithms to enable the agent to analyze user interactions via vision and audio, providing contextually relevant assistance in voice
  • Work closely with research scientists and engineers, contributing to team discussions, sharing knowledge, and actively participating in code reviews to foster a collaborative environment
  • Proactively identify and address technical challenges, stay updated on the latest AI advancements, and focus on developing solutions that can be effectively integrated into Google products and services, contributing to product impact

Skills

LLMs
Multimodal AI
Vision Processing
Audio Processing
Text Processing
LLM Post-Training
LLM Evaluation
AI Agents

DeepMind

Develops artificial general intelligence systems

About DeepMind

This company leads in the field of artificial general intelligence (AGI), with notable applications across healthcare, energy management, and biotechnology. Their work in early diagnostic tools for eye diseases, optimizing energy usage in major data centers, and groundbreaking contributions to protein structure prediction underlines their commitment to harnessing AI for diverse practical applications. The company's dedication to pushing the boundaries of AI technology not only propels the industry forward but also creates a dynamic and impactful working environment for its employees.

London, United KingdomHeadquarters
2010Year Founded
$4.9MTotal Funding
ACQUISITIONCompany Stage
AI & Machine Learning, BiotechnologyIndustries
1,001-5,000Employees

Benefits

Performance Bonus

Risks

Emerging AI models may challenge DeepMind's current strategies.
Backlash against AI models like Gemini poses reputational risks.
Labeling AI-generated content could increase operational complexity for DeepMind.

Differentiation

DeepMind combines AI, ML, and neuroscience for general-purpose learning algorithms.
DeepMind's AlphaFold model advances protein folding research significantly.
GraphCast by DeepMind offers rapid, accurate ten-day weather forecasts.

Upsides

AI-driven drug discovery is set to grow significantly in 2024.
AlphaCode 2 showcases AI's potential in competitive programming.
DeepMind's AI tools are transforming music creation and meteorology.

Land your dream remote job 3x faster with AI