Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I am a passionate software engineering specialist currently focused on Reinforcement Learning from Human Feedback (RLHF) at Outlier AI. My work involves guiding AI agents through complex coding tasks and refining their generated code to significantly improve model capabilities and debugging accuracy. I thrive on solving challenging problems and enhancing AI-driven systems to produce effective solutions in coding and UI/UX design. With a strong background in machine learning and software development, I have successfully driven performance optimization in real-time video processing and model acceleration projects. I also enjoy supporting others in their learning journeys as a teaching assistant. My expertise spans multiple programming languages and frameworks, and I am dedicated to delivering high-quality, efficient, and innovative software solutions.…I am a passionate software engineering specialist currently focused on Reinforcement Learning from Human Feedback (RLHF) at Outlier AI. My work involves guiding AI agents through complex coding tasks and refining their generated code to significantly improve model capabilities and debugging accuracy. I thrive on solving challenging problems and enhancing AI-driven systems to produce effective solutions in coding and UI/UX design. With a strong background in machine learning and software development, I have successfully driven performance optimization in real-time video processing and model acceleration projects. I also enjoy supporting others in their learning journeys as a teaching assistant. My expertise spans multiple programming languages and frameworks, and I am dedicated to delivering high-quality, efficient, and innovative software solutions.

Mitansh Desai

Full Stack Developer, AI Prompt Engineer, AI Engineer, +2





I am a passionate software engineering specialist currently focused on Reinforcement Learning from Human Feedback (RLHF) at Outlier AI. My work involves guiding AI agents through complex coding tasks and refining their generated code to significantly improve model capabilities and debugging accuracy. I thrive on solving challenging problems and enhancing AI-driven systems to produce effective solutions in coding and UI/UX design. With a strong background in machine learning and software development, I have successfully driven performance optimization in real-time video processing and model acceleration projects. I also enjoy supporting others in their learning journeys as a teaching assistant. My expertise spans multiple programming languages and frameworks, and I am dedicated to delivering high-quality, efficient, and innovative software solutions.…I am a passionate software engineering specialist currently focused on Reinforcement Learning from Human Feedback (RLHF) at Outlier AI. My work involves guiding AI agents through complex coding tasks and refining their generated code to significantly improve model capabilities and debugging accuracy. I thrive on solving challenging problems and enhancing AI-driven systems to produce effective solutions in coding and UI/UX design. With a strong background in machine learning and software development, I have successfully driven performance optimization in real-time video processing and model acceleration projects. I also enjoy supporting others in their learning journeys as a teaching assistant. My expertise spans multiple programming languages and frameworks, and I am dedicated to delivering high-quality, efficient, and innovative software solutions.

Available to hire

I am a passionate software engineering specialist currently focused on Reinforcement Learning from Human Feedback (RLHF) at Outlier AI. My work involves guiding AI agents through complex coding tasks and refining their generated code to significantly improve model capabilities and debugging accuracy. I thrive on solving challenging problems and enhancing AI-driven systems to produce effective solutions in coding and UI/UX design.

With a strong background in machine learning and software development, I have successfully driven performance optimization in real-time video processing and model acceleration projects. I also enjoy supporting others in their learning journeys as a teaching assistant. My expertise spans multiple programming languages and frameworks, and I am dedicated to delivering high-quality, efficient, and innovative software solutions.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Language

Javanese

Intermediate

Work Experience

Software Engineering Specialist for RLHF at Outlier AI

November 1, 2024 - Present

Guided AI agents through complex coding tasks, evaluating and refining generated code to boost model capabilities. Refined AI agent’s Chain of Thought (COT) reasoning to achieve ideal PR fixes in 25+ SWE Bench GitHub issues, significantly enhancing model-generated code and debugging accuracy. Contributed to the training of agentic models for UI/UX design, optimizing COT reasoning and generated code to produce 20+ ideal UI/UX solutions and measurably enhance design quality.

Machine Learning Engineer at Ross Video

April 30, 2024 - July 21, 2025

Transitioned preprocessing from CPU to GPU-based using Torchvision, reducing data prep time by 52%. Engineered real-time video super resolution with GANs, optimizing for GPU efficiency and cost-effectiveness. Refactored two legacy PyTorch 0.3 custom convolution layers (deformable, dynamic local) to PyTorch 2.0 with ATen in C++, resulting in improvements in the deinterlacer model. Accelerated super-resolution, segmentation, and deinterlacer models with TensorRT and int8 quantization, achieving up to 85% latency reduction and real-time performance while maintaining output quality. Optimized image preprocessing augmentation using CUDA, reducing execution time by 36%.

Teaching Assistant at Carleton University

April 30, 2024 - July 21, 2025

Led office hours for Functional Programming (Haskell) and Cognitive Science courses across 4 terms. Assessed term papers and quizzes for classes of up to 300 students, providing constructive feedback to support student development.

Full Stack Developer at Ericsson

December 31, 2022 - July 21, 2025

Refactored UI of internal tools with React’s component architecture and state management, increasing user retention by 18% for development teams. Designed visualization solutions using Kibana to monitor code coverage and Jenkins job performance, improving build times by ∼7 minutes and increasing quality of repository commits. Integrated Elasticsearch with secure REST API endpoints by implementing certificate-based authentication, enabling reliable and protected data retrieval for dashboard visualizations and monitoring tools.

Software Developer at Canadian Resident Matching Service (CaRMS)

December 31, 2021 - July 21, 2025

Remodeled the CaRMS Match Day landing page using the Bootstrap framework, increasing page load speed by 28%. Developed a report monitoring page to track and build administrative functions for a set of user application reports.

Software Engineering Specialist for RLHF at Outlier AI

November 1, 2024 - Present

Guided AI agents through complex coding tasks, evaluating and refining generated code to boost model capabilities. Refined the Chain of Thought (COT) reasoning of AI agents to achieve ideal pull request fixes in 25+ software engineering benchmark GitHub issues, enhancing code generation and debugging accuracy. Contributed to training agentic models for UI/UX design with optimized reasoning and code generation that produced over 20 ideal UI/UX solutions, significantly improving design quality.

Machine Learning Engineer at Ross Video

April 30, 2024 - August 21, 2025

Transitioned preprocessing workflows from CPU to GPU-based implementations using Torchvision, reducing data preparation time by 52%. Engineered real-time video super resolution solutions using GANs optimized for GPU performance and cost-efficiency. Refactored legacy PyTorch 0.3 custom convolution layers to PyTorch 2.0 using ATen in C++, enhancing the deinterlacer model's performance. Achieved up to 85% latency reduction and real-time performance in super-resolution, segmentation, and deinterlacer models using TensorRT and int8 quantization, while maintaining output quality. Optimized image preprocessing augmentation via CUDA, reducing execution time by 36%.

Teaching Assistant at Carleton University

April 30, 2024 - August 21, 2025

Led office hours for Functional Programming (Haskell) and Cognitive Science courses over four academic terms. Assessed term papers and quizzes for classes of up to 300 students, providing constructive feedback to foster student development and learning.

Full Stack Developer at Ericsson

December 31, 2022 - August 21, 2025

Refactored UI of internal tools using React's component architecture and state management, leading to an 18% increase in user retention among development teams. Designed visualization solutions with Kibana for monitoring code coverage and Jenkins job performance, improving build times by approximately 7 minutes and enhancing commit quality. Integrated Elasticsearch with secure REST API endpoints through certificate-based authentication, ensuring reliable and secure data retrieval for dashboards and monitoring tools.

Software Developer at Canadian Resident Matching Service (CaRMS)

December 31, 2021 - August 21, 2025

Redesigned the CaRMS Match Day landing page using Bootstrap, increasing page load speed by 28%. Developed a report monitoring page to manage and build administrative functions for user application reports.