We are sharing a specialised consulting opportunity for experienced Senior Software Engineers with strong expertise in Python, Java, Rust, C++, Go, TypeScript, algorithms, debugging, feature development, codebase refactoring, and performance optimisation to contribute to an advanced AI training and reinforcement-learning environment project using Model Context Protocol (MCP) tools.
Selected professionals will design reproducible reinforcement-learning environments that test advanced AI systems on realistic software-engineering problems requiring both strong coding ability and effective interaction with MCP servers. Tasks may involve bug fixing, feature implementation, refactoring, performance optimisation, information discovery, deterministic verification, and development of golden reference solutions. No prior experience in AI is required.
Key Responsibilities
Reinforcement Learning Environment Development
-
Design realistic software-engineering environments for AI evaluation
-
Create tasks involving bug fixing, feature development, refactoring, and performance optimisation
-
Build reproducible environments with clearly defined expected behaviour
-
Develop technical scenarios that require structured reasoning over unfamiliar systems
-
Ensure tasks accurately measure both software-engineering ability and tool use
Model Context Protocol (MCP) Integration
-
Design tasks requiring agents to interact with real MCP servers
-
Create workflows that require discovery and retrieval of relevant information through MCP tools
-
Evaluate whether agents use available tools appropriately and efficiently
-
Build scenarios combining external context with software-engineering problem solving
-
Define expected MCP interaction patterns where required for task completion
Software Development & Debugging
-
Work across C++, Python, Java, Go, TypeScript, Rust, or comparable languages
-
Diagnose complex defects and identify root causes
-
Implement maintainable fixes rather than superficial patches
-
Analyse unfamiliar codebases and dependencies
-
Apply strong software-engineering practices throughout task development
Feature Development & Refactoring
-
Design and implement realistic feature-development scenarios
-
Refactor existing code to improve clarity, maintainability, and architecture
-
Preserve intended functionality while improving implementation quality
-
Evaluate alternative design and implementation approaches
-
Ensure solutions remain scalable and maintainable
Performance Optimisation
-
Develop tasks involving software-performance bottlenecks
-
Analyse algorithmic and implementation-level inefficiencies
-
Improve latency, throughput, or resource utilisation where appropriate
-
Validate that optimisations preserve correctness
-
Document performance reasoning and technical trade-offs clearly
Verification & Golden Reference Solutions
-
Develop deterministic verification mechanisms for each environment
-
Create golden reference solutions demonstrating expected task completion
-
Define objective acceptance criteria for AI-generated outputs
-
Ensure tests reliably distinguish correct from incomplete or incorrect solutions
-
Validate both tool usage and software-engineering outcomes
Code Review & Technical Quality
-
Review technical submissions for correctness, maintainability, and clarity
-
Identify hidden defects, unsupported assumptions, or incomplete implementations
-
Apply rigorous software-engineering standards to evaluation tasks
-
Contribute to development of consistent technical-quality criteria
-
Incorporate review feedback into evolving environment designs
Ideal Profile
-
Strong professional software-engineering experience
-
Proficiency in one or more of C++, Python, Java, Go, TypeScript, or Rust
-
Deep understanding of algorithms and data structures
-
Strong performance-tuning and optimisation skills
-
Demonstrated experience debugging complex software systems
-
Strong background in feature implementation and codebase refactoring
-
Proven ability to develop maintainable and scalable solutions
-
Experience working with large or distributed codebases is highly valuable
-
Strong code-review experience and familiarity with software best practices
-
Excellent written and verbal technical communication skills
-
Comfortable collaborating across cross-functional and remote teams
-
Familiarity with modern AI or machine-learning systems is advantageous but not required
-
No prior AI-training experience is required
Engagement Details
-
Part-time independent contractor engagement
-
Fully remote
-
Expected commitment: approximately 15 hours per week
-
Displayed compensation range: $80–$120/hour
-
Actual compensation structure is output-based, with payment made per task that meets project specifications
-
Task completion time may vary depending on individual experience and workflow
-
Minimum weekly submission requirements apply; the source does not specify the exact number of required tasks
-
Work will involve reinforcement-learning environment creation, MCP tool use, debugging, feature implementation, refactoring, performance optimisation, deterministic verification, and golden reference solutions
-
Screening includes an approximately 30-minute AI interview, with a technical assessment potentially included before hiring-manager review
-
Roles are typically filled within approximately 48 hours
-
Selected experts are expected to begin initial tasks within approximately 24–48 hours after onboarding
-
Project scope, workload, technical environments, and evaluation standards may evolve depending on project requirements
-
Work must be completed without using confidential or proprietary information belonging to any employer, client, institution, or other third party
About the Platform
This opportunity is available through 24-MAG LLC. We connect experienced professionals with remote consulting opportunities across technical, evaluation, and project-based workstreams.
By submitting this application, you acknowledge that your information may be processed by 24-MAG LLC for recruitment and opportunity matching in accordance with our Privacy Policy: https://www.24-mag.com/privacy-policy
