PatSnap

Senior/ Machine Learning Engineer (LLM)

Singapore

Not SpecifiedCompensation
Senior (5 to 8 years)Experience Level
Full TimeJob Type
UnknownVisa
AI, AI & Machine Learning, Natural Language ProcessingIndustries

Requirements

Candidates should possess a Master's or Ph.D. in Computer Science, AI, or a related field, along with 5 or more years of experience in machine learning with a focus on NLP and Large Language Models (LLMs). They should have a strong understanding of transformer architectures and modern LLM frameworks such as BERT, GPT, and T5, and demonstrate proficiency in deep learning frameworks like PyTorch, TensorFlow, and JAX. Experience with distributed training systems like Megatron and optimization techniques is also preferred.

Responsibilities

The Senior/Machine Learning Engineer (LLM) will engage in the development and optimization of large-scale pre-training language models, including model architecture design, parallel training strategies, and performance improvements. They will drive research and implementation of advanced LLM post-training techniques, such as chain-of-thought tuning and preference alignment. The role involves developing and optimizing data collection pipelines, designing and implementing model deployment solutions, and collaborating with cross-functional teams to apply LLM capabilities in various business scenarios, while staying current with the latest advancements in the field and contributing to the company’s technical roadmap.

Skills

Machine Learning
NLP
LLMs
Transformer architectures
PyTorch
TensorFlow
JAX
Python
Hugging Face
DeepSpeed
Distributed training
Megatron
Pre-training
Model architecture design
Parallel training
Inference optimization
BERT
GPT
T5

PatSnap

Intellectual property and innovation intelligence platform

About PatSnap

PatSnap offers a platform that helps businesses, inventors, and researchers understand patents and innovation. Its main product aggregates and analyzes data from patents, scientific literature, and market reports, enabling clients to make informed decisions about their R&D investments. PatSnap operates on a subscription model, providing various service tiers and educational courses to empower clients in leveraging their innovation data. The company's goal is to help clients drive business growth and maintain a competitive edge through effective use of intellectual property.

Singapore, SingaporeHeadquarters
2007Year Founded
$342MTotal Funding
SERIES_ECompany Stage
Data & Analytics, Consulting, EducationIndustries
501-1,000Employees

Benefits

Paid Vacation
Unlimited Paid Time Off
Mental Health Support
Parental Leave
Life Insurance
Private Healthcare
Private Pension
RRSP Contribution Matching
Eyecare Voucher Scheme
WFH Stipend
Cycle To Work scheme
Company Equity
Paid Volunteering Days

Risks

UK lawtech startups' growth could threaten PatSnap's market share by 2026.
Cultural resistance to legal tech may limit PatSnap's effectiveness in some companies.
Expansion in Japan may face challenges due to current minimal revenue share.

Differentiation

PatSnap integrates generative AI into R&D, enhancing drug discovery timelines.
The platform aggregates data from patents, literature, and market reports for comprehensive insights.
PatSnap's subscription model ensures steady revenue and continuous platform enhancement.

Upsides

Expansion in Japan aims to boost revenue contribution to 10% from a minimal share.
Collaboration with BizInt Smart Charts enhances data integration and patent intelligence access.
CoPilot AI assistant improves IP and R&D workflows, increasing user efficiency.

Land your dream remote job 3x faster with AI