Job Closed

This listing is no longer active.

Innodata logo
Innodata

Innodata, with over 35 years of expertise, is a trusted leader in data solutions and AI innovation. The company specializes in training and deploying generative

Applied Research Scientist, LLM Evaluation – Post-Training

Location

New Jersey

Posted

104 days ago

Salary

0

Seniority

Senior

Postgraduate Degree5 yrs expEnglishPythonPyTorchTensorFlow

Job Description

Applied Research Scientist, LLM Evaluation – Post-Training

Innodata

• Define and execute a research agenda focused on LLM evaluation and post-training, especially evaluation-driven model improvement • Design rigorous experiments to study how evaluation methodologies impact fine-tuning and post-training outcomes • Develop and validate evaluation frameworks for LLM and multimodal systems, including: benchmark/task design scoring methods judge/model-assisted evaluation human evaluation protocols robustness/stress testing • Lead research on advanced evaluation domains, including long-context, cross-modal, and dynamic multi-turn evaluations • Study the effectiveness and limitations of existing evaluation techniques, and propose improved methodologies with clear validity and scalability tradeoffs • Analyze model behavior and failure patterns; generate actionable recommendations for model improvement and evaluation redesign • Collaborate with AI/ML Research Engineers to translate research methods into scalable evaluation and post-training pipelines • Collaborate with Language Data Scientists to integrate human-in-the-loop and synthetic data/evaluation strategies into research programs • Engage with customer technical stakeholders to understand evaluation goals, review methodologies, and provide expert recommendations • Contribute to internal benchmark datasets, evaluation frameworks, and reusable research assets • Produce high-quality technical documentation, internal research reports, and client-facing materials explaining methods, results, assumptions, and limitations • Contribute to thought leadership and best practices in LLM evaluation, post-training, and GenAI quality measurement

Job Requirements

  • MS/PhD in Computer Science, Machine Learning, Statistics, Applied Mathematics, AI, or a related quantitative scientific field (PhD strongly preferred)
  • 5+ years of relevant experience in applied research / research science in ML/AI, with substantial work in LLMs or foundation models
  • Demonstrated experience with LLM evaluation, benchmarking, alignment, post-training, or model quality research
  • Strong foundation in experimental design, statistical analysis, and scientific reasoning for ML systems
  • Strong coding skills in Python for research experimentation and analysis (e.g., data processing, evaluation pipelines, statistical analysis, visualization)
  • Experience working with modern ML tooling/frameworks (e.g., PyTorch, Hugging Face, JAX/TensorFlow as applicable)
  • Ability to evaluate and compare human and automated evaluation methods, including tradeoffs in cost, reliability, validity, and scalability
  • Experience designing evaluation studies and protocols that are reproducible across datasets, model versions, and evaluation runs
  • Ability to collaborate directly with technical stakeholders including research scientists, ML engineers, data scientists, and customer technical counterparts
  • Strong communication skills and ability to present nuanced technical conclusions, assumptions, and limitations clearly.

Benefits

  • Health insurance
  • Retirement plans
  • Paid time off
  • Flexible work arrangements
  • Professional development opportunities

Related Categories

Related Job Pages

More Research Scientist Jobs

Mindrift logo

Evaluation Scenario Writer - AI Agent Testing Specialist

Mindrift

Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid. Project time expectations: Tasks are estimated to require around 10–20 hours per week during active phases, based on project requirements; This is an estimate, not a guaranteed workload, and applies only while the project is active. Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Research Scientist105 days ago

Please submit your CV in English and indicate your level of English proficiency. Mindrift connects specialists with project-based AI opportunities for leading tech companies, focused on testing, evaluating, and improving AI systems. Participation isproject-based, not permanent employment. What this opportunity involves You’ll create challenging coding test cases that push AI coding systems to their limits: - Review and refine realistic coding tasks based on provided production codebases with realistic scope, requirements and information sources  - Write comprehensive functional tests that validate actual end-to-end behavior and edge-cases, not just superficial checks - Craft “fair but hard” challenges where the AI has all the context it needs, but has to work for it (information scattered across files and external sources, complex reasoning required) - Analyze AI failures to understand what the model struggles with vs. what it masters - Iterate based on feedback from expert QA reviewers who score your work on 7 quality criteria What we look for This opportunity is a good fit for experienced developers, software engineers, and/or test automation specialists open to part-time, non-permanent projects. Ideally, contributors will have:  - Degree in Computer Science, Software Engineering or related fields - 5+ years in software development, primarily Python (pytest, async/await, subprocess, file operations) - Background in Full-Stack development, with an equal focus on building React-based interfaces and robust Back-end systems - Experience writing tests (functional, integration – not just running them) - Docker containers (running evaluations locally in containers) - CI/CD understanding (GitHub Actions as a user: triggers, labels, reading results) - English proficiency - B2 How it works Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid Effort estimate  Tasks for this project are estimated to take 20 hours to complete, depending on complexity. This is an estimate and not a schedule requirement; you choose when and how to work. Tasks must be submitted by the deadline and meet the listed acceptance criteria to be accepted. Payment - Paid contributions, with rates up to $80/hour* - Fixed project rate or individual rates, depending on the project - Some projects include incentive payments *Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Iowa
Job Closed
Mindrift logo

Evaluation Scenario Writer - AI Agent Testing Specialist

Mindrift

Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid. Project time expectations: Tasks are estimated to require around 10–20 hours per week during active phases, based on project requirements; This is an estimate, not a guaranteed workload, and applies only while the project is active. Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Research Scientist105 days ago

Please submit your CV in English and indicate your level of English proficiency. Mindrift connects specialists with project-based AI opportunities for leading tech companies, focused on testing, evaluating, and improving AI systems. Participation isproject-based, not permanent employment. What this opportunity involves You’ll create challenging coding test cases that push AI coding systems to their limits: - Review and refine realistic coding tasks based on provided production codebases with realistic scope, requirements and information sources  - Write comprehensive functional tests that validate actual end-to-end behavior and edge-cases, not just superficial checks - Craft “fair but hard” challenges where the AI has all the context it needs, but has to work for it (information scattered across files and external sources, complex reasoning required) - Analyze AI failures to understand what the model struggles with vs. what it masters - Iterate based on feedback from expert QA reviewers who score your work on 7 quality criteria What we look for This opportunity is a good fit for experienced developers, software engineers, and/or test automation specialists open to part-time, non-permanent projects. Ideally, contributors will have:  - Degree in Computer Science, Software Engineering or related fields - 5+ years in software development, primarily Python (pytest, async/await, subprocess, file operations) - Background in Full-Stack development, with an equal focus on building React-based interfaces and robust Back-end systems - Experience writing tests (functional, integration – not just running them) - Docker containers (running evaluations locally in containers) - CI/CD understanding (GitHub Actions as a user: triggers, labels, reading results) - English proficiency - B2 How it works Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid Effort estimate  Tasks for this project are estimated to take 20 hours to complete, depending on complexity. This is an estimate and not a schedule requirement; you choose when and how to work. Tasks must be submitted by the deadline and meet the listed acceptance criteria to be accepted. Payment - Paid contributions, with rates up to $80/hour* - Fixed project rate or individual rates, depending on the project - Some projects include incentive payments *Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Florida
Job Closed
Mindrift logo

Evaluation Scenario Writer - AI Agent Testing Specialist

Mindrift

Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid. Project time expectations: Tasks are estimated to require around 10–20 hours per week during active phases, based on project requirements; This is an estimate, not a guaranteed workload, and applies only while the project is active. Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Research Scientist105 days ago

Please submit your CV in English and indicate your level of English proficiency. Mindrift connects specialists with project-based AI opportunities for leading tech companies, focused on testing, evaluating, and improving AI systems. Participation isproject-based, not permanent employment. What this opportunity involves You’ll create challenging coding test cases that push AI coding systems to their limits: - Review and refine realistic coding tasks based on provided production codebases with realistic scope, requirements and information sources  - Write comprehensive functional tests that validate actual end-to-end behavior and edge-cases, not just superficial checks - Craft “fair but hard” challenges where the AI has all the context it needs, but has to work for it (information scattered across files and external sources, complex reasoning required) - Analyze AI failures to understand what the model struggles with vs. what it masters - Iterate based on feedback from expert QA reviewers who score your work on 7 quality criteria What we look for This opportunity is a good fit for experienced developers, software engineers, and/or test automation specialists open to part-time, non-permanent projects. Ideally, contributors will have:  - Degree in Computer Science, Software Engineering or related fields - 5+ years in software development, primarily Python (pytest, async/await, subprocess, file operations) - Background in Full-Stack development, with an equal focus on building React-based interfaces and robust Back-end systems - Experience writing tests (functional, integration – not just running them) - Docker containers (running evaluations locally in containers) - CI/CD understanding (GitHub Actions as a user: triggers, labels, reading results) - English proficiency - B2 How it works Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid Effort estimate  Tasks for this project are estimated to take 20 hours to complete, depending on complexity. This is an estimate and not a schedule requirement; you choose when and how to work. Tasks must be submitted by the deadline and meet the listed acceptance criteria to be accepted. Payment - Paid contributions, with rates up to $80/hour* - Fixed project rate or individual rates, depending on the project - Some projects include incentive payments *Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Colorado
Job Closed
Mindrift logo

Evaluation Scenario Writer - AI Agent Testing Specialist

Mindrift

Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid. Project time expectations: Tasks are estimated to require around 10–20 hours per week during active phases, based on project requirements; This is an estimate, not a guaranteed workload, and applies only while the project is active. Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Research Scientist105 days ago

Please submit your CV in English and indicate your level of English proficiency. Mindrift connects specialists with project-based AI opportunities for leading tech companies, focused on testing, evaluating, and improving AI systems. Participation isproject-based, not permanent employment. What this opportunity involves You’ll create challenging coding test cases that push AI coding systems to their limits: - Review and refine realistic coding tasks based on provided production codebases with realistic scope, requirements and information sources  - Write comprehensive functional tests that validate actual end-to-end behavior and edge-cases, not just superficial checks - Craft “fair but hard” challenges where the AI has all the context it needs, but has to work for it (information scattered across files and external sources, complex reasoning required) - Analyze AI failures to understand what the model struggles with vs. what it masters - Iterate based on feedback from expert QA reviewers who score your work on 7 quality criteria What we look for This opportunity is a good fit for experienced developers, software engineers, and/or test automation specialists open to part-time, non-permanent projects. Ideally, contributors will have:  - Degree in Computer Science, Software Engineering or related fields - 5+ years in software development, primarily Python (pytest, async/await, subprocess, file operations) - Background in Full-Stack development, with an equal focus on building React-based interfaces and robust Back-end systems - Experience writing tests (functional, integration – not just running them) - Docker containers (running evaluations locally in containers) - CI/CD understanding (GitHub Actions as a user: triggers, labels, reading results) - English proficiency - B2 How it works Apply → Pass qualification(s) → Join a project → Complete tasks → Get paid Effort estimate  Tasks for this project are estimated to take 20 hours to complete, depending on complexity. This is an estimate and not a schedule requirement; you choose when and how to work. Tasks must be submitted by the deadline and meet the listed acceptance criteria to be accepted. Payment - Paid contributions, with rates up to $80/hour* - Fixed project rate or individual rates, depending on the project - Some projects include incentive payments *Note: Rates vary based on expertise, skills assessment, location, project needs, and other factors. Higher rates may be offered to highly specialized experts. Lower rates may apply during onboarding or non-core project phases. Payment details are shared per project.

Alabama
Job Closed