Data Freelancers Wanted! Join Us at Devoteam Göteborg

This part-time, remote contract invites Python Data Scraping Engineers based in Sweden to support the Tendem project at toloka-ai. The position focuses on building and managing data scraping workflows within a hybrid AI and human system. As an AI Pilot, collaboration with Tendem Agents is central, with particular attention to quality control, critical thinking, and data accuracy. Key responsibilities Oversee end-to-end data extraction from complex websites, ensuring structured datasets are accurate and reliable. Utilize internal tools such as Apify and OpenRouter, along with custom workflows, to streamline data collection, validation, and task execution. Adapt extraction strategies for dynamic and interactive web sources, including those with JavaScript-rendered content and evolving site structures. Maintain high data quality through validation checks, consistency controls, formatting standards, and systematic verification before delivery. Scale scraping operations for large datasets using batching or parallelization, monitor for failures, and adjust to site changes to ensure stability. Requirements Minimum 3 years of experience in data engineering, web scraping, automation, or similar fields. Proficiency in Python and other relevant programming languages. Strong understanding of web technologies and data structures. Hands-on experience with a variety of scraping tools and techniques. Ability to troubleshoot and optimize data extraction workflows. Strong problem-solving skills and attention to detail. Compensation Contributors can earn up to $32 per hour, depending on expertise and pace. Actual compensation may vary by project scope, complexity, and required skills. How to apply Submit your application as directed by Mindrift. Qualified applicants will be matched with projects that fit their technical background and can work on their own schedule. Projects may include coding, automation, and refining AI outputs, all contributing to the advancement of AI in real-world settings.

Please submit your CV in English and specify your English proficiency level. toloka-ai, in partnership with Mindrift, offers project-based contract opportunities for experienced data science engineers. Projects focus on testing, evaluating, and improving AI systems for leading technology clients. This is not a permanent employment role but rather contract work based on individual assignments. Role overview Freelance Data Science Engineers (Python & SQL) work on a variety of projects, each with unique analytical and computational challenges. Assignments involve designing original data science problems that reflect real-world business scenarios in sectors such as telecom, finance, government, e-commerce, and healthcare. Create computational problems that require advanced analytical workflows and cannot be solved manually in a short time. Design tasks solvable in Python, using libraries like Pandas, Numpy, Scipy, scikit-learn, Statsmodels, Matplotlib, and Seaborn. Develop scenarios involving data processing, statistical analysis, feature engineering, predictive modeling, and extracting business insights. Ensure each problem is deterministic and reproducible, using fixed seeds when randomness is necessary. Base challenges on real business needs, such as customer analytics, risk assessment, fraud detection, forecasting, optimization, and operational efficiency. Cover the full data science workflow: from data ingestion and cleaning to modeling, validation, and deployment considerations. Include big data scenarios that require scalable computational solutions. Validate solutions in Python with established libraries and statistical methods. Document each problem clearly, providing realistic business context and verified answers. Requirements Minimum 5 years of hands-on data science experience with measurable business outcomes. Portfolio of completed projects or publications showing practical problem-solving skills. Advanced Python programming for data science, including experience with pandas, numpy, scipy, scikit-learn, and statsmodels. Strong knowledge of statistical analysis and machine learning algorithms. Proficient SQL skills and experience with database operations for analysis. Familiarity with Generative AI tools and techniques such as LLMs, RAG, prompt engineering, and vector databases. Understanding of MLOps and model deployment workflows. Knowledge of frameworks like TensorFlow, PyTorch, or LangChain. Excellent written English at C1 level or above. Application process Submit your CV and indicate your English proficiency level. Complete the qualification steps. If selected, join a project. Work on assigned tasks. Receive compensation for completed work. Location: Remote , Sweden

Please submit your CV in English and include your English proficiency level. Toloka AI, in partnership with Mindrift, offers project-based freelance opportunities for skilled professionals to help test, evaluate, and improve AI systems for leading technology companies. This is a freelance, project-based position and does not lead to permanent employment. Role overview As a Freelance AI Trainer (Data Scientist) working remotely from Sweden, assignments focus on building and refining AI models through real-world data science challenges. Typical projects include: Designing computational data science problems that mirror analytical workflows in sectors like telecom, finance, government, e-commerce, and healthcare. Developing problems that require Python solutions, using libraries such as Pandas, Numpy, Scipy, Scikit-learn, Statsmodels, Matplotlib, and Seaborn. Ensuring problems are computationally demanding, taking days or weeks to solve, and not just manual calculations. Creating scenarios that call for complex reasoning in data processing, statistical analysis, feature engineering, predictive modeling, and drawing insights. Making sure all challenges are deterministic and reproducible, either by avoiding randomness or by using fixed random seeds. Basing tasks on real business needs, including customer analytics, risk assessment, fraud detection, forecasting, optimization, and operational efficiency. Covering the full data science workflow: data ingestion, cleaning, exploratory analysis, modeling, validation, and deployment considerations. Including big data processing scenarios that need scalable computational approaches. Validating solutions in Python with standard data science libraries and statistical techniques. Documenting problem statements clearly, with realistic business contexts and verified solutions. Requirements Minimum 5 years of hands-on data science experience with demonstrated business impact. A portfolio of completed projects or publications showing real-world problem solving. Advanced Python skills for data science (including pandas, numpy, scipy, scikit-learn, statsmodels). Strong background in statistical analysis and machine learning, with deep knowledge of algorithms and their uses. Proficiency in SQL and database operations for data manipulation and analysis. Experience with Generative AI, such as LLMs, Retrieval-Augmented Generation, prompt engineering, and vector databases. Understanding of MLOps and model deployment processes. Familiarity with frameworks like TensorFlow, PyTorch, and LangChain. Excellent written English (C1 level or higher). How to apply Apply Complete qualifications Join a project Fulfill tasks Receive payment

Please submit your CV in English and include your English proficiency level. Mindrift connects experienced professionals with project-based AI roles at top technology companies. Assignments focus on testing, evaluating, and improving AI systems. All work is freelance and project-based, not permanent employment. Role overview This freelance role is designed for optical engineers with strong Python skills who want to contribute to AI projects. Assignments vary, and each brings new technical challenges. Typical tasks include: Writing original computational physics problems based on real research workflows Developing programming challenges that require Python solutions, using libraries such as Numpy, SciPy, or Sympy Designing computationally intensive problems that cannot be solved manually in a reasonable timeframe Creating scenarios involving complex reasoning in areas like mechanics, electromagnetism, thermodynamics, and quantum mechanics Basing problems on real-world research or practical physics applications Using Python and standard physics simulation libraries to verify solutions Documenting problem statements clearly and providing accurate, verified solutions Requirements Degree in Physics (theoretical, experimental, or computational) or a related discipline Proficiency in Python for numerical validation; experience with MATLAB, R, C, SQL, Numpy, Pandas, or SciPy is a plus At least 2 years of professional experience in applied, research, or teaching roles Understanding of numerical simulation methods Ability to design problems that reflect real physics research workflows Creative approach to problem design across physics fields Familiarity with physics modeling and approximation techniques Strong written English skills at C1 level or above Project commitment Assignments usually require 10–20 hours per week during active project phases. Actual workload may vary depending on project needs and is not guaranteed. Compensation Contributors can earn up to $45 per hour, depending on project complexity, scope, and individual contribution. Compensation may vary by project. Application process Apply Complete the qualification process Join a project Finish assigned tasks Receive compensation Location: Remote, Sweden

Please submit your CV in English and include your English proficiency level. Mindrift connects specialized professionals with project-based AI assignments for leading technology companies. The focus is on testing, evaluating, and improving AI systems. All participation is project-based and does not imply permanent employment. Role overview This freelance position involves developing and validating computational engineering problems that help train AI systems. Projects vary, and contributors may work on tasks such as: Creating original computational engineering problems that mirror real-world workflows Designing problems that require Python programming for engineering calculations or simulations Ensuring problems are computationally intensive, using numerical methods or iterative solutions Developing scenarios related to system design, optimization, and analysis Basing problems on actual research challenges or practical engineering issues Validating solutions in Python, using standard engineering libraries Documenting problem statements clearly and providing verified, correct answers Requirements Degree in Electrical Engineering or a related field Proficiency in Python for numerical validation; familiarity with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or similar libraries is valued At least 2 years of professional experience in applied, research, or teaching roles Understanding of practical engineering constraints and approximations Ability to write clearly in English at C1 level or higher Professional certifications (such as CMME, SAS, CAP) and experience with international or applied projects are a plus Project workflow The process follows these steps: Application → Qualification Process → Project Engagement → Task Completion → Payment. Time commitment During active project phases, tasks typically require 10–20 hours per week. Actual workload may vary depending on project needs. Compensation Contributors may earn up to $50 per hour, depending on expertise and pace. Rates can vary by project based on complexity, scope, and required skills. Location This remote, project-based role is open to candidates based in Sweden.

Toloka AI seeks a freelance AI trainer who combines civil engineering experience with strong Python programming skills. This contract position is fully remote and open to candidates based in Sweden. Role overview This role centers on applying civil engineering knowledge and Python expertise to help train and refine artificial intelligence systems. The work supports the development of AI applications by contributing specialized insights from both fields. What you will do Apply civil engineering concepts and Python programming to improve AI models Offer subject matter expertise to shape and support AI application development Location This is a remote position for candidates located in Sweden.

Toloka AI seeks a Mechanical Engineer with advanced Python skills for a freelance AI Trainer role. This contract position is fully remote and open to candidates based in Sweden. Key responsibilities Create training materials that focus on AI technologies Lead workshops designed for engineers and technical teams Collaborate with fellow engineers to enhance AI products Team collaboration This role works closely with Toloka AI's engineering group. The environment emphasizes practical experience and clear, effective communication to support product development and knowledge sharing.

Please submit your CV in English and specify your English proficiency level. This freelance, project-based position at Mindrift connects material scientists with expertise in Python to AI-driven projects for leading technology companies. The work focuses on testing, evaluating, and improving AI systems. This is not a permanent role. Role overview Material scientists in this role design and validate advanced engineering problems for AI training. Typical tasks include: Creating materials engineering problems that mirror real-world workflows Developing challenges that use Python for engineering calculations and simulations Ensuring problems are computationally intensive, often using numerical methods or iterative solutions Designing problems related to system design, optimization, and analytical processes Grounding tasks in real research or practical engineering scenarios Validating solutions in Python, often with established engineering libraries Documenting problem statements and providing verified solutions Requirements Degree in Material Science or a related discipline Strong Python skills for numerical validation; experience with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or similar libraries is also valued Minimum of 2 years of professional experience in applied work, research, or teaching Solid understanding of practical engineering constraints and approximations Excellent written English at C1 level or higher How to get started Apply Meet qualification criteria Join a project Complete assigned tasks Receive payment Time commitment During active project phases, work typically requires 10–20 hours per week. Workload may vary and is not guaranteed outside of active projects. Compensation Contributors can earn up to $45 per hour, depending on project scope, complexity, and expertise. Rates may differ across projects on the platform. Location: Remote , Sweden

Mindrift connects mathematics specialists with project-based AI work for leading technology companies. This freelance role focuses on testing, evaluating, and improving AI systems. Projects are temporary and do not represent permanent employment. Role overview Assignments vary and may include: Designing original computational mathematics problems that reflect real research workflows Formulating problems that require Python programming for solutions, often using libraries like Numpy, SciPy, or Sympy Creating computationally intensive problems that cannot be solved manually in a reasonable timeframe Developing challenges involving advanced reasoning in number theory, combinatorics, graph theory, or numerical analysis Basing problems on authentic research challenges or practical applications in mathematics Validating solutions with Python and standard mathematical libraries Documenting problem statements clearly and providing verified solutions Requirements Degree in Mathematics (Pure or Applied) or a related discipline Strong Python skills for numerical validation; experience with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or other domain-specific libraries is also valued At least 2 years of experience in applied mathematics, research, or teaching Background in numerical methods and symbolic computation Ability to design problems that simulate real mathematical research processes Familiarity with computational complexity theory Excellent written English (C1+ proficiency) Project commitment Workload is estimated at 10–20 hours per week while projects are active. Actual hours may vary depending on project needs and are not guaranteed outside active phases. Compensation Contributors may earn up to $45 per hour, depending on expertise and contribution speed. Rates can vary by project based on scope and complexity. Other projects on the platform may offer different earning potentials. Application process Apply Pass qualifications Join a project Complete assigned tasks Receive payment Please submit your CV in English and specify your English proficiency level.

Toloka AI is looking for a freelance specialist with a strong background in both biology and Python programming. This remote role is based in Sweden and centers on developing and improving materials used to train artificial intelligence systems. Role overview This position focuses on creating and refining training content that combines biological science concepts with Python. The work supports AI development by ensuring materials are accurate and effective. What you will do Develop AI training materials that draw on biology and Python expertise Review and enhance existing content to improve clarity and accuracy Who this role suits This freelance opportunity is well suited to those interested in the overlap between biological sciences and artificial intelligence, and who enjoy working independently on specialized projects.

Please submit your CV in English and indicate your English proficiency level. Mindrift connects specialists with project-based AI work for leading technology companies. Projects focus on testing, evaluating, and improving AI systems. All participation is on a freelance, project-by-project basis and does not constitute permanent employment. Role overview This freelance position is designed for physicists with Python experience who are interested in part-time, non-permanent projects. Each project presents unique challenges, and contributors play a key role in shaping computational physics tasks. Work is fully remote and based in Sweden. Design computational physics problems that reflect real-world research workflows Create tasks that require Python programming solutions, often using libraries such as Numpy, SciPy, or Sympy Ensure problems are computationally intensive, not solvable manually within days or weeks Develop problems involving complex reasoning in fields like mechanics, electromagnetism, thermodynamics, and quantum mechanics Base tasks on real research challenges or practical physics applications Verify solutions in Python, using standard physics simulation libraries Document problem statements clearly and provide validated solutions Requirements Degree in Physics (theoretical, experimental, or computational) or a related discipline Proficiency in Python for numerical validation; familiarity with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, Stata, or other programming languages is also valued Minimum of 2 years of professional experience in applied fields, research, or teaching Experience with numerical simulation methods Ability to design problems that reflect real physics research workflows Creative problem-solving skills across multiple physics domains Knowledge of physics modeling and approximation techniques Strong written English skills at C1+ level Application process Apply Pass qualifications Join a project Complete tasks Receive payment Time commitment Tasks typically require about 10–20 hours per week during active project phases. Workload may fluctuate and is not guaranteed outside of these periods. Compensation Earn up to $45 per hour, depending on expertise and pace. Compensation varies by project based on scope, complexity, and required skills. Other projects on the platform may offer different rates depending on their needs. Location: Remote , Sweden

We Want to Collaborate with You!We are on the lookout for experienced data consultants who are eager to co-create tomorrow's digital solutions with us. Our offices are located in Malmö, Helsingborg, Halmstad, Göteborg, and Stockholm, collaborating with some of the most exciting companies in the region.More than just a company, we aspire to be a partner in your success!With 25 years of global experience and over 10,000 employees across 20 countries, we stand as a leading partner for AWS, Azure, Snowflake, and Google Cloud, providing expertise and innovation in every project. Our focus at Devoteam Göteborg is primarily on helping Nordic companies harness their data through software, modern data platforms, advanced analytics, and AI-driven solutions.

Please submit your CV in English and indicate your English proficiency level. Mindrift connects specialists with project-based AI work for leading technology companies. This freelance position focuses on testing, evaluating, and improving AI systems. The role is project-based and not a permanent placement. Role overview As a Freelance AI Trainer (Research Physicist with Python Skills), you will design and validate computational physics problems to support AI development. The work is remote within Sweden and organized around specific projects. What you will do Create original computational physics problems that reflect real-world research workflows Formulate problems that require Python programming for solutions, using libraries such as NumPy, SciPy, and SymPy Ensure problems are computationally intensive and not solvable manually within a reasonable timeframe Develop challenges involving complex reasoning in mechanics, electromagnetism, thermodynamics, and quantum mechanics Design problems inspired by actual research or practical applications in physics Validate solutions using Python and standard physics simulation libraries Document problem statements clearly and provide verified solutions Requirements Degree in Physics (theoretical, experimental, or computational) or a related field Proficiency in Python for numerical validation; experience with MATLAB, R, C, SQL, NumPy, Pandas, SciPy, or similar tools is valued At least 2 years of relevant professional experience (applied, research, or teaching) Experience with numerical simulation methods Ability to design problems that mirror authentic physics research workflows Creative problem-solving skills across multiple physics domains Understanding of physics modeling and approximation techniques Strong written English skills (C1 level or higher) How projects work The process includes: application, passing qualifications, joining a project, completing tasks, and receiving payment. Time commitment During active project phases, expect roughly 10–20 hours per week, depending on project needs. Actual workload may vary. Compensation Earn up to $45 per hour, depending on contribution and project scope. Compensation varies by project based on complexity and required expertise. Other projects on the platform may offer different pay structures.

Please submit your CV in English and include your English proficiency level. Mindrift Technologies matches subject matter experts with project-based AI work for major technology clients. Assignments focus on testing, evaluating, and improving AI systems. This is a project-based, non-permanent position. Role overview The Freelance Quantum Research Scientist & AI Trainer (Python) works on a variety of short-term projects. Responsibilities shift by assignment and may include: Designing computational physics problems that reflect real-world research workflows Developing problems that require Python programming to solve, using libraries such as Numpy, SciPy, and Sympy Creating computationally intensive tasks that are not feasible to solve manually in a reasonable time Formulating scenarios that involve complex reasoning in mechanics, electromagnetism, thermodynamics, and quantum mechanics Basing problems on authentic research challenges or practical applications in physics Validating solutions in Python with standard simulation libraries Documenting problem statements and providing verified, correct answers Preferred qualifications Degree in Physics (theoretical, experimental, or computational) or a closely related field Strong Python programming skills for numerical validation; experience with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or similar libraries is valued At least 2 years of professional experience in applied, research, or teaching roles Background in numerical simulation methods Experience designing problems that mirror real physics research processes Creative problem-solving across physics domains Knowledge of physics modeling and approximation techniques Fluent written English at C1 level or higher How projects work Apply Pass qualification(s) Join a project Complete assigned tasks Receive payment Time commitment During active project periods, tasks usually require about 10–20 hours per week. Actual workload depends on project needs. Compensation Earn up to $45 per hour, depending on expertise and task completion rate. Pay rates may vary by project scope and required skills. Other projects on the platform may offer different compensation levels. This is a remote position based in Sweden.

Please submit your CV in English and indicate your English proficiency level. toloka-ai, through Mindrift, offers freelance, project-based work for specialists interested in evaluating and improving AI systems. This contract role is remote and based in Sweden. The position is not a permanent employment arrangement, and work is structured around specific projects for leading technology companies. Role overview This freelance AI Trainer position focuses on developing and validating advanced computational statistics problems. Work centers on creating original tasks inspired by real mathematical research, often requiring Python programming and the use of libraries such as Numpy, SciPy, or Sympy. Projects may involve computationally intensive problems that take significant time to solve, with an emphasis on areas like number theory, combinatorics, graph theory, and numerical analysis. What you will do Create original computational statistics problems modeled on authentic research workflows Develop problems that require Python-based solutions, using libraries such as Numpy, SciPy, or Sympy Design tasks that may take days or weeks to solve due to their complexity Write problem statements involving advanced reasoning in mathematics and numerical analysis Base tasks on real research challenges or practical mathematical applications Validate all solutions using Python and established mathematical libraries Document each problem clearly and provide verified, accurate answers Requirements Degree in Statistics or a related field Proficiency in Python for numerical validation (experience with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or similar libraries is also considered) Minimum of 2 years of professional experience, which can include applied research or teaching Strong written English skills at C1 level or higher Professional certifications (such as CMME, SAS Certifications, CAP) and experience with international or applied projects are a plus How to join Application Qualification Assessment Project Participation Task Completion Compensation Project details During active project periods, tasks typically require 10-20 hours per week. Actual workload may vary and is not guaranteed outside of active project phases. Compensation Earnings can reach up to $50 per hour, depending on expertise and pace. Compensation rates may vary by project, based on complexity, scope, and required skills.

Toloka Annotators is looking for a Freelance AI Trainer - Data Annotation Specialist to help build better AI models by producing well-labeled datasets. This remote position is open to candidates based in Sweden. Key responsibilities Label and annotate various types of data for AI training and evaluation tasks Support the accuracy and performance of machine learning models through careful data preparation Work remotely with an international team of collaborators What we value Keen attention to detail when handling data Interest in AI technologies and maintaining data quality Ability to work independently from a home office Residence in Sweden Work arrangement This freelance role offers the flexibility to work from anywhere within Sweden. Team members connect and collaborate online with colleagues worldwide.

Join our dynamic team as an AI Data Specialist and contribute to the enhancement of AI-generated content in Swedish. Job Type: FreelanceLocation: Work from homeWork Schedule: Part-time - 10+ hours per week. Enjoy the flexibility to work at your convenience.Start Date: Immediate availabilityDuration: To be confirmedRate: Compensation varies based on location. Contribute to the Future of AIAre you a student, recent graduate, stay-at-home parent, gig worker, or professional seeking flexible remote opportunities? If you're passionate about shaping the evolution and safety of contemporary AI models, we want to hear from you! Your ResponsibilitiesIn this role, you will perform a variety of data-centric tasks, including:Data collection, evaluation, and annotation.Conducting pairwise comparisons.Executing counting tasks.Tagging and labeling objects across various content types (audio, video, images, or collected data). QualificationsNative-level fluency in SwedishProficient in English (Fluent or advanced proficiency, levels B2–C2)Experience in AI & Data (Preferred): Familiarity with machine learning tasks, data collection and preprocessing, quality assurance, and data annotation. What We ProvideFlexible working hoursPotential to earn additional incomePrompt paymentIdeal for students, part-timers, or stay-at-home parents

Mindrift offers project-based remote roles for specialists in chemistry and Python, supporting the development and evaluation of AI systems for technology companies. This freelance position is open to candidates based in Sweden and is not a permanent employment opportunity. Role overview As a Freelance AI Trainer, the main focus is to create and validate computational chemistry problems that mirror real-world research workflows. Tasks include designing original scenarios that require advanced reasoning in physical chemistry, quantum chemistry, and molecular modeling. Problems should demand Python programming for resolution, often using libraries such as numpy, scipy, or other relevant chemical packages. Each problem must be computationally intensive, with solutions that cannot be completed manually in a reasonable timeframe. All problem statements and answers should be clearly documented and verified using established computational chemistry methods in Python. Requirements Degree in Chemistry or a related field Proficiency in Python for numerical validation (experience with MATLAB, R, C, SQL, Numpy, Pandas, SciPy, or domain-specific libraries is also acceptable) At least 2 years of professional experience in applied, research, or teaching roles Experience with numerical methods in chemistry Strong understanding of computational chemistry concepts and techniques Excellent written English skills at C1+ level Project commitment During active project phases, expect to contribute approximately 10–20 hours per week. Actual workload may vary depending on project needs. Compensation Contributors can earn up to $45 per hour, depending on expertise and the pace of contributions. Compensation may differ between projects based on complexity, scope, and required skills. How to apply Please submit your CV in English and include your English proficiency level. The process involves applying, passing qualifications, joining a project, completing assigned tasks, and receiving compensation.

Please submit your CV in English and include your English proficiency level. Project Overview Toloka AI, in partnership with Mindrift, offers project-based freelance roles focused on testing and evaluating AI systems. As a Freelance AI Evaluation Engineer, the main responsibility is to build and assess datasets that measure how well AI coding agents perform on tasks similar to those faced by real-world developers. This is a contract-based position and does not lead to permanent employment. Key Responsibilities Design virtual companies from broad outlines, creating codebases, infrastructure, and realistic supporting materials such as documentation, tickets, and internal communications to simulate development history. Develop and refine tasks that represent intermediate milestones within these virtual companies. Define prompts, set evaluation standards, and ensure tasks are both solvable and fairly judged. Create assignments in isolated environments that mirror a developer's workstation. This setup includes a Linux machine with development tools, MCP servers for repositories, task tracking, messaging, documentation, and a working web application codebase. Write tests that accept all valid solutions and reject incorrect ones, maintaining a fair but rigorous standard. Collaborate with AI agents to ensure tests surface real problems, filter out poor solutions, and confirm strong results. Review code generated by AI agents, analyze success or failure reasons, and design edge cases or adversarial scenarios to challenge model capabilities. Apply feedback from expert QA reviewers who check your work against established quality benchmarks. What This Role Does Not Include Data labeling Prompt engineering Writing code from scratch (the AI agent handles most coding; your focus is on guidance and evaluation) This work relies heavily on collaboration with AI systems. Building tasks that truly test advanced models requires direct interaction with these agents and thoughtful evaluation of their outputs.