LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness) | $100-$120/hr Remote
Overview
We're seeking an experienced machine learning researcher to tackle open-ended empirical problems at the intersection of vision, language, and model robustness. In this remote, project-based role, you'll train and refine deep learning systems end-to-end, often under tight constraints on data, compute, and model size. Work spans image classification, generative image models, LLM alignment, and adversarial resilience — with a strong emphasis on practical, measurable results.
What You'll Do5
- 1Train image classifiers and generative image models from the ground up, and fine-tune open-weight language models for specific behaviors.
- 2Get maximum performance from limited datasets, compute budgets, and small model sizes.
- 3Harden models against both adversarial inputs and adversarial conversational tactics.
- 4Compress models via techniques like pruning, quantization, and distillation to meet strict size and latency limits while preserving accuracy.
- 5Troubleshoot and resolve training instabilities or convergence issues as they arise.
Requirements7
- 13+ years of hands-on machine learning research experience (including PhD research) with deep expertise in at least one of: adversarial robustness, efficient computer vision, generative image modeling, LLM post-training, or multilingual pre-training.
- 2Advanced proficiency with PyTorch, JAX, TensorFlow, or another major deep learning framework, and a track record of training models end-to-end.
- 3Solid understanding of adversarial robustness methods such as PGD-based adversarial training, TRADES, and evaluation with AutoAttack under L∞ threat models, including avoiding gradient masking.
- 4Experience with model compression techniques (quantization, pruning, knowledge distillation) and deploying models under hard size or latency constraints on edge or embedded devices.
- 5For LLM work: hands-on experience with SFT, preference optimization (DPO, RLHF, RLAIF), synthetic dataset construction, and steering multi-turn conversational behavior.
- 6For multilingual work: experience training tokenizers and models for low-resource or typologically diverse languages, and balancing skewed per-language data.
- 7A degree from a top-100 university, experience at a FAANG or similar AI organization, or a strong publication/open-source record.
Who Should Apply
You're a researcher who likes digging into messy, empirical ML problems and iterating until the model behaves. You're comfortable working independently on loosely specified research questions, and you care about making models not just accurate but robust, efficient, and deployable. You have strong coding skills in modern ML frameworks and can point to concrete projects — papers, open-source work, or deployed systems — that show your ability to train and improve deep learning models end-to-end.
Salary Insight
The role pays $100–$120 per hour on a project basis, with flexible, remote work and the opportunity to collaborate with leading AI researchers.
Location
Required Skills
Application Tip
In your application, walk through one specific end-to-end training project: describe the problem, the model architecture, your data/compute constraints, and the exact techniques you used to improve robustness, efficiency, or sample quality — include links to code or papers if available.
See NearSkill jobs more often in your search
How your application is processed
1Application received
Your resume and details are logged the moment you apply.
2ATS + eligibility screening
We check your profile against the role’s skills, seniority, and requirements.
3Employer sees qualified profiles only
Only candidates who clear screening move forward.
Similar open positions
Explore active roles that match your skills and interests.
Turing
VerifiedLLM Java Developer
This role centers on building and maintaining Java back-end components that support LLM training and refinement. Daily work includes running model evaluations, scoring AI responses, and writing clear rationales for those judgments. You will also contribute to Supervised Fine-Tuning (SFT) datasets and collaborate on Reinforcement Learning with Human Feedback (RLHF). The position is a fully remote contractor assignment with a required 4-hour overlap with PST and a commitment of 20, 30, or 40 hours per week.
Turing
VerifiedSenior LLM Engineer
A remote role based in India centers on designing and building Generative AI and LLM systems with Python and Langchain. The engineer will create RAG pipelines, prompt techniques, and agent-based workflows that run in production. The position expects 7-12 years of experience and close work with engineering teams, business SMEs, and data teams to shape the LLM roadmap. Strong SQL and cloud familiarity across AWS, Azure, or GCP support the day-to-day work.
Micro1
VerifiedForward Deployed Engineer
This role sits at the intersection of applied AI, ML infrastructure, and partner-facing product development. You'll work directly with leading AI labs and enterprises to transform ambiguous research questions into production-grade systems. As a Member of Technical Staff, you'll own everything from data curation and LLM agent workflows to deployment and partner success — all while operating in a fast-moving, remote-first environment.
Turing
VerifiedLLM C/ C++ Developer
A company building next-generation dialog agents for education, entertainment, and question-answering needs a C++ engineer to review and validate AI-generated code. In this contract role, you will debug C/C++ code produced by AI systems, help define new features with cross-functional teams, and contribute to training LLM back-end components. You will also improve public GitHub repositories and mentor other developers through code review. The position is remote and contract-based, with flexible hours and a two-step interview process that includes a 60-minute technical session.
Mercor
VerifiedLegal Domain Expert — AI Training & Evaluation
A senior legal practitioner sits inside a frontier AI lab's research team, where professional judgment becomes training signal for models that reason about legal work. The day-to-day involves reviewing the quality of legal tasks and model outputs, writing instruction specifications and golden solutions, and designing evaluation benchmarks. The engagement is a full-time W-2 position through Cincinnatus LLC, and the work takes place on-site in the Bay Area multiple days per week. This role needs a practicing specialist with deep domain knowledge, not a generalist or an academic.
Micro1
VerifiedAI Jailbreak & Prompt-Injection Security Expert
micro1 is looking for a contractor to join a forward-thinking client project focused on hardening AI systems against exploitation. Your mission: design creative adversarial tests—think ethical jailbreaks, prompt injection, and tool-use abuse—to uncover weaknesses in modern LLMs. No deep AI background is required; your hands-on security expertise and adversarial mindset are what count.