Robert Kimani

Key Skills

Software

Appen

Top Subject Matter

Agent-based AI systems

function calling

LLM workflows

Top Data Types

Text

Document

Computer Code Programming

Top Task Types

Function Calling

Fine-tuning

Computer Programming/Coding

Freelancer Overview

I am a Python-focused Software Engineer with hands-on experience(~2 years) building, validating, and optimizing AI-driven systems, including LLM workflows, agentic architectures, and data pipelines. I have a track record of contributing to global AI initiatives through contract roles with leading platforms, delivering high-quality code, structured datasets, and scalable validation systems. I bring a strong expertise in Python, SQL, JSON workflows, and AI-assisted development environments, with a focus on accuracy, performance, and real-world applicability of AI systems.

Labeling Experience

LLM Agentic Trainer (JSON / SQL Workflows) at Turing

As an LLM Agentic Trainer, I built and validated structured datasets to support agent-based AI systems using JSON and SQL workflows. I ensured accuracy, consistency, and alignment with real-world use cases within large-scale training environments. I also designed robust task structures and produced key documentation for best dataset development practices. • Developed and maintained datasets tailored for agent-function workflows. • Focused on ensuring function-calling reliability in AI models using structured data. • Collaborated with teams to standardize validation processes. • Used Python and SQL-based workflows throughout the labeling process.

As an LLM Agentic Trainer, I built and validated structured datasets to support agent-based AI systems using JSON and SQL workflows. I ensured accuracy, consistency, and alignment with real-world use cases within large-scale training environments. I also designed robust task structures and produced key documentation for best dataset development practices. • Developed and maintained datasets tailored for agent-function workflows. • Focused on ensuring function-calling reliability in AI models using structured data. • Collaborated with teams to standardize validation processes. • Used Python and SQL-based workflows throughout the labeling process.

2025 - 2026

Agent Function Calling LLM Trainer at Turing

Function Calling

As an Agent Function Calling LLM Trainer, I was responsible for designing conversational datasets for AI assistants and implementing function calling logic. I utilized JSON schemas to simulate real-world interactions and sharpened model reasoning abilities. My work focused on evaluating AI assistant performance in complex scenarios and refining their tool usage. • Created multi-turn, structured datasets for conversational function-calling. • Evaluated and optimized model interaction reliability. • Enhanced context comprehension in LLMs through meticulous testing. • Developed standards for iterative AI function-calling labeling.

As an Agent Function Calling LLM Trainer, I was responsible for designing conversational datasets for AI assistants and implementing function calling logic. I utilized JSON schemas to simulate real-world interactions and sharpened model reasoning abilities. My work focused on evaluating AI assistant performance in complex scenarios and refining their tool usage. • Created multi-turn, structured datasets for conversational function-calling. • Evaluated and optimized model interaction reliability. • Enhanced context comprehension in LLMs through meticulous testing. • Developed standards for iterative AI function-calling labeling.

2025 - 2025

AI Technical Integrity & Quality Control at Appen

Appen

In this quality control role, I developed validation and auditing tools to safeguard the technical integrity of AI-generated code and related datasets. Python scripts were leveraged to automate code cleanup, enforce coding standards, and perform structural checks. The role demanded attention to detail in standardizing production environments and quality assurance. • Designed validation scripts for notebook/code pipeline review. • Automated processes for consistent dataset quality. • Focused on technical and structural dataset integrity. • Standardized file systems for robust AI code management.

In this quality control role, I developed validation and auditing tools to safeguard the technical integrity of AI-generated code and related datasets. Python scripts were leveraged to automate code cleanup, enforce coding standards, and perform structural checks. The role demanded attention to detail in standardizing production environments and quality assurance. • Designed validation scripts for notebook/code pipeline review. • Automated processes for consistent dataset quality. • Focused on technical and structural dataset integrity. • Standardized file systems for robust AI code management.

2024 - 2025

Education

Work History