LLM Research Scientist – Pre-training, Computer Vision, Adversarial Robustness
Employment Information
Report this job
Job expired or something wrong with this job?
Job Description
LLM research scientist training and improving frontier AI models for vision, language, and adversarial robustness. Conducting empirical research and optimizing models under data, compute, size, and latency constraints.
Responsibilities:
- Train image classifiers and generative image models from scratch
- Fine-tune open-weight language models
- Solve empirical, open-ended machine learning research problems
- Optimize models within limited data, compute, and model-size budgets
- Improve model robustness against adversarial inputs and conversations
- Compress models to meet hard size and latency constraints without sacrificing accuracy
- Diagnose and resolve training issues
- Collaborate with leading AI researchers on projects training and enhancing frontier AI systems
Requirements:
- 3+ years of machine learning research experience; PhD research counts toward this requirement
- Strong experience with PyTorch, JAX, TensorFlow, or similar ML frameworks
- Degree from a top-100 university, experience at a FAANG or comparable AI company, or an equivalent research track record through publications or impactful open-source contributions
- Expertise in one or more of adversarial robustness, efficient computer vision, generative image modeling, LLM post-training and behavioral robustness, or multilingual pre-training
- Experience with adversarial training of image classifiers, robust-accuracy evaluation, and managing robustness–accuracy trade-offs
- Experience training image classifiers end-to-end, model compression, and deployment under hard size or latency budgets
- Experience training image generative models from scratch and evaluating sample quality using metrics such as FID
- Experience with supervised fine-tuning and preference optimization of open-weight language models
- Experience shaping conversational behavior over multiple turns and preserving general capability during behavior-specific fine-tuning
- Experience training multilingual or low-resource-language models from scratch and designing tokenizers across scripts
- Additional experience in scaling laws, curriculum learning, model evaluation, uncertainty estimation, data augmentation, or synthetic data is a plus
- Must not require H-1B or STEM OPT support
Benefits:
- Flexible, project-based work
- Competitive compensation
- Fully remote work
- Work on your own schedule
- Weekly payments via Stripe or Wise
- Collaborate with leading AI researchers
- Projects can be extended, shortened, or concluded early depending on needs and performance
- Reasonable accommodations upon request
- Referral bonuses of up to $480 per successful referral
Similar Jobs
LLM Research Scientist – Pre-training, Computer Vision, Adversarial Robustness
LLM research scientist training and improving vision and language models for Mercor’s AI lab and enterprise partners. Focusing on adversarial robustness, computer vision, generative modeling, and multilingual pre-training.
Research Scientist – First Author
Research Scientist validating AI-generated reproductions, simulations, and results for an AI-for-science benchmark. Contributing to a public paper and evaluation set as lead or first author.
Research Scientist – First Author
Research Fellow validating AI-generated scientific research for a specialized AI lab. Creating benchmark tasks, evaluation standards, and a public first-author research paper.
Research Scientist – First Author
Research Fellow validating AI-generated scientific research for a specialized AI lab. Creating benchmark tasks, evaluation standards, and a public first-author research paper.
Research Assistant – AI Development
Research Assistant developing AI conversational systems and prototypes for Johns Hopkins education research. Managing study data, APIs, user-activity tracking, and iterative feature testing.
Research Scientist Intern
Research Scientist Intern advancing Pluralis’s decentralized large-model training research. Publishing foundational Protocol Learning papers toward frontier-scale models.
Master's/Ph.D. Research Fellow – Credit Risk Modeling, Python, Statistics
Pesquisador bolsista desenvolvendo modelos de rating e provisão de crédito para o Itaú. Aplicação de Python, estatística e análise de dados em projeto de inovação.
Cybersecurity Research Assistant
Cybersecurity Research Assistant securing Kubernetes and HPC infrastructure for Axle’s NIH rare-disease research. Building controls, monitoring, automation, and AI-assisted security tooling.
Principal Scientist – Ecological Services
Principal Scientist leading EHS Support’s ecological consulting practice. Developing business, managing teams, and delivering ecological risk assessments, NRDA, restoration, and monitoring services.
Senior Applied Research Scientist, Data Curation
Senior Applied Research Scientist developing NVIDIA’s petabyte-scale multimodal data-curation and deduplication pipelines. Advancing document extraction for foundation-model training and NVIDIA Nemotron.
Senior/Principal Scientist, Formulation and Process Development
Formulation scientist leading drug product and process development for Celcuity’s targeted oncology therapies. Supporting formulations, scale-up, lifecycle management, and regulatory filings.
Mercor Research Fellowship – APEX
Research fellow building and validating APEX benchmarks for Mercor, an AI data and enterprise evaluation company. Designing evaluations, testing frontier models, and publishing results.
Research Assistant I
Part-time Research Assistant translating and disseminating UAMS health research findings. Developing evidence-based websites, reports, presentations, and community-facing materials.
Research Assistant – Infectious Diseases
Part-time remote research assistant supporting WashU’s infectious-disease and global public-health research. Conducting assays, analyzing data, coordinating studies, and supporting project operations.
Senior Research Scientist, Information Science – Extended Temporary
Senior Research Scientist developing AI/NLP systems for scientific feasibility assessment at the University of Arizona. Leading reproducible research pipelines, empirical studies, publications, and student collaboration.
Senior Researcher, Mixed Methods
Senior Researcher advancing decarbonization through behavioral and mixed-methods research at GoodPower. Designing experiments, evaluating interventions, and translating evidence into stronger energy-transition programs.
Senior Applied Scientist – Cyber Defense
Senior Applied Scientist building and deploying AI agents for NVIDIA’s cybersecurity operations. Evaluating models, optimizing agent systems, and automating detection, investigation, and response.
Senior Applied Research Scientist – Computational Geometry, Meshing
NVIDIA applied researcher developing computational geometry, meshing, and discretization algorithms. Advancing GPU-based CAD-to-simulation workflows across engineering, semiconductor, and scientific computing.
Research Scientist, PhD New Grad
Research Scientist building and deploying machine-learning models for SentiLink’s identity verification and fraud-risk products. Researching fraud patterns, engineering features, and delivering production code and analyses.
Applied Scientist, PhD New Grad
Applied Scientist building production fraud-detection models for SentiLink’s identity-risk solutions. Researching fraud, engineering features, and deploying real-time machine-learning systems.