Hemanth Kokkiligadda

Proof of Experience

Mercor · Jul 2026 – Aug 2026

Evaluated AI-generated responses for accuracy, relevance, reasoning quality, instruction-following, and overall response quality.
Created and refined prompts for LLMs to test reasoning, problem-solving, factual accuracy, and adherence to task requirements.
Reviewed and annotated text, image, and multimodal AI outputs using detailed evaluation guidelines and scoring rubrics.
Identified hallucinations, logical errors, inconsistencies, and unsupported claims, providing clear feedback to improve model performance.
Performed quality assurance and comparative evaluation of model responses to ensure consistency, accuracy, and alignment with project requirements.
Applied prompt engineering techniques including few-shot prompting, structured instructions, and prompt chaining for Al evaluation and training tasks.
Collaborated with project teams to maintain annotation standards, resolve edge cases, and deliver high-quality AI training data.

Turing · May 2026 – Aug 2026

Evaluated and improved AI-generated responses across text, image, video, and multimodal tasks based on detailed quality guidelines.
Designed, reviewed, and refined prompts to improve model accuracy, relevance, reasoning, and instruction-following.
Performed data annotation, response evaluation, quality assurance, and error analysis to identify hallucinations, inconsistencies, and policy or instruction violations.
Applied prompt engineering techniques such as few-shot prompting, prompt chaining, and structured instructions to assess and improve LLM performance.
Conducted comparative evaluations of AI outputs using predefined rubrics and provided actionable feedback for model improvement.
Collaborated with AI training and QA teams to maintain annotation consistency, meet project requirements, and deliver high-quality training data.

Deccan AI · Jun 2025 – May 2026

Experienced AI Trainer with expertise in data annotation, prompt engineering, and AI model evaluation across image, video, text, and response-based projects.
Worked on improving NLP, computer vision, and generative AI models by creating high-quality datasets, reviewing outputs, and optimizing prompt performance.
Collaborated with data scientists, QA teams, and analysts to enhance model accuracy, reduce errors, and ensure ethical, context-aware responses.
Performed quality assurance (QA) to maintain accuracy and consistency in project deliverables while ensuring workflow efficiency.
Continuously analyzed performance metrics and suggested improvements for data collection, labeling, and model tuning processes.
Worked as a panellist/recruiter, evaluating candidates and coordinating with teams to support hiring and project quality standards.

Amazon · Jun 2024 – Jun 2025

Assisted Amazon's third-party sellers in navigating and succeeding in the Amazon marketplaces in the EU region.
Troubleshot issues such as investigating and resolving problems including listing errors, policy violations, and performance metrics.
Escalated complex cases when the root cause was not found using the tools provided, collaborating with internal teams to ensure timely and effective resolution.
Managed client relationships, organizing regular meetings to discuss ongoing projects and future needs.

Nxtwave Disruptive · Sep 2022 – Mar 2024

Completed daily list of assignments, contributing to engineer team by handling basic tasks and learning new skills.
Contributed insights during weekly team meetings to enhance project outcomes.
Worked with production engineers to update documentation in line with production techniques.
Maintained logbook over duration of training period, recording projects undertaken and skills learned.
Engaged in continuous learning through professional development courses, staying abreast of industry advancements.
Analysed technical specifications and requirements, aiding in the precise execution of engineering projects.

No completed sessions yet. Anything Hemanth finishes shows up here with the evidence behind it.

J.B Institute of Engineering and Technology · 2018 — 2022