LLM Research Consultant

Sony Research India is driving cutting-edge research and development in various locations around the globe, including laboratories in Japan, the United States, Europe, and Asia. We endeavor to create new technology, products, and services while sustaining Sony Group's diverse businesses in electronics, entertainment, and financial fields. For our research Centre to blaze a trail in the latest technologies, we seek to foster the growth of a diverse pool of research and engineering talent and create a technology talent bank to drive research excellence worldwide. Sony Research India is offering outstanding career opportunities around frontline technologies such as AI and data analytics.

What we are looking for:

We are seeking a highly motivated LLM Research Consultant with deep expertise in large language model architectures and their application to diverse downstream tasks, including Neural Machine Translation (NMT), Natural Language Processing (NLP), and Natural Language Understanding (NLU). The ideal candidate should possess strong knowledge of state-of-the-art LLM design principles and training paradigms, and demonstrate hands-on experience in developing and extending advanced model architectures for multilingual and cross-domain settings.

The role requires expertise in modern LLM techniques such as advanced inference optimization, model steering and controllability, mechanistic interpretability, mixture-of-experts (MoE) architectures, learning dynamics analysis, structured reasoning approaches (e.g., tree- and graph-based reasoning methods), and scalable training and acceleration strategies. The candidate should be well-versed in current research trends and capable of adapting cutting-edge LLM methodologies to improve performance on complex real-world tasks.

The consultant will collaborate closely with research scientists and engineers to design scalable, efficient, and high-quality multilingual systems, particularly for entertainment-domain translation and related language technologies across Indian and international languages.

Key Responsibilities:

• Help in design and development of LLM-based production systems for multilingual applications, including NMT, NLP, and NLU, ensuring scalability, reliability, and deployment readiness.
• Cutting-edge LLM research (e.g., controllable generation, reasoning frameworks, MoE architectures, inference optimization) into practical product features that improve quality, efficiency, and user experience.
• Optimize end-to-end pipelines for domain-specific use cases, particularly entertainment-domain dialogue and subtitle translation, balancing model performance with latency and cost constraints.
• Implement and evaluate advanced inference strategies, including decoding optimization, model steering, memory-efficient fine-tuning, and throughput acceleration techniques suitable for real-world deployment.
• Collaborate closely with engineering teams to integrate models into APIs, backend services, and production infrastructure.
• Contribute to applied research efforts by experimenting with emerging LLM techniques and assessing their impact on product performance and differentiation.
• Provide technical assistance in benchmarking, experimentation, and scaling strategies to ensure high-quality, robust multilingual systems.

Work Location:

Remote within India

Duration of the paid one year contract:

• The annual paid direct contractual tenure is extendable.
• Ideally this position will start from April first week of 2026.
• The working hours are from 9:00 to 18:00 (Monday to Friday) full-time.

Essential Education:

• Ph.D. or Masters in Computer Science background or Bachelor's with hands-on experience on Python with an applications in Language Processing.

Skills required:

Essential Skills:

• Strong background in Deep Learning and Natural Language Processing
• Excellent programming skills in Python, with practical experience in PyTorch
• Solid understanding of LLMs and foundation models.
• Ability to read, implement, and extend research papers.

Good to have Skills:

• Strong mathematical and algorithmic reasoning skills.
• Publications or prior research exposure in top-tier conferences/journals (e.g., ICLR, ICML, NeurIPS, ACL, EMNLP, AAAI).
• Experience with large-scale experiments, or production-grade NLP systems.

Our Values:

Dreams & Curiosity: Pioneer the future with dreams and curiosity.

Diversity: Pursue the creation of the very best by harnessing diversity and varying viewpoints.

Integrity & Sincerity: Earn the trust for Sony brand through ethical and responsible conduct.

Sustainability: Fulfil our stakeholder responsibilities through disciplined business practices.

Sony Research India is committed to equal opportunity in all its employment practices, policies and procedures and to ensuring that no worker or potential worker will receive less favourable treatment due to any characteristic protected under applicable local laws.