Uppu Raghu

Software Engineering graduate with hands-on experience working as an AI Contractor at Mercor on advanced AI agent training and evaluation projects. Contributed to Imperium and Compact OTS projects focused on building intelligent AI systems capable of handling complex coding, browsing, and terminal workflows. Skilled in prompt engineering, rubric creation, markdown documentation, technical writing, and AI workflow evaluation with a strong foundation in software engineering and autonomous AI systems. - Working as an AI contractor on advanced AI agent training and evaluation projects focused on coding, browsing, and terminal-based environments. * Contributed to Imperium and Project Compact OTS initiatives for building long-horizon AI agents capable of handling complex workflows across software engineering tasks. * Created and refined high-quality training datasets, markdown (.md) documentation, workflow instructions, and rubric-based evaluation frameworks for AI systems. * Worked as an SWE Writer and Software Engineering Writer, designing structured prompts, technical reasoning tasks, and agent evaluation content. * Developed compact conversation summaries and context-preservation workflows to improve AI memory handling and autonomous task execution. * Assisted in rubric creation, response validation, quality assessment, and benchmark generation for AI coding agents. * Built strong experience in AI data annotation, prompt engineering, technical writing, workflow analysis, and software engineering evaluation. * Collaborated in projects involving complex coding environments, browser-based workflows, and terminal interaction simulations to improve next-generation AI agents. * Strengthened expertise in Markdown documentation, AI workflow optimization, software engineering concepts, and scalable evaluation pipelines. * Gained hands-on exposure to modern AI development ecosystems focused on autonomous coding assistants and intelligent agent systems.

1. Researched and analyzed multi-agent Al workflows for enterprise code generation, extracting key insights on agent roles, workflow types, evaluation metrics, and implementation strategies. 2. Designed and implemented a proof of concept using agentic workflows to automate enterprise-level code generation by analyzing problem statements, identifying relevant code from GitHub repositories, and generating or modifying code to deliver complete solutions. 3. Created and implemented an agentic workflow leveraging task tree traversal and LLMs (Claude) to generate and refine code based on MVCTs. This approach optimized chat history usage by 60% by passing only the necessary content to the LLM for code generation and modification, ensuring efficiency within token limits. Proficient in crafting high-impact prompts for ChatGPT and Claude to perform tasks like code generation, summarization, data extraction, and complex logic simulation. • Designed and implemented intelligent agent workflows that interact with LLMs to analyze problem statements, retrieve relevant GitHub code, and generate or adapt solutions. • Optimized token usage and enhanced response precision by using task-tree traversal techniques and modular prompt strategies. • Hands-on experience working with OpenAI’s GPT models, Anthropic Claude, and tools like LangChain, Pydantic, Streamlit, Flask, and Postman for full-stack AI app development.

Worked as a Freelance AI Data Annotator and ML Labeling Specialist on multiple AI/ML projects involving data annotation, quality validation, and model training support. Contributed to improving machine learning datasets through accurate labeling and review processes across different project domains.

I am happy to share that I’m starting a new position as a web developer intern at Nucleus Technology through a training-oriented internship program in Nullclass.