Swmkwr Narzary

Experienced Team Lead and Reviewer for Frontier AI Lab projects, with expertise in designing complex multi-step evaluation rubrics, stress-testing AI models, and driving quality assurance processes. Skilled in crowd management, auditing workflows, and developing high-impact learning resources. Proven track record as Quality Analyst and Project Lead across leading AI companies.

Team Lead / Reviewer / Rubrics Expert at Mercor (2025-09 – Present)

Review work submitted by experts, provide feedback, and approve it for client delivery. Document edge cases and assist in improving learning materials. Create complex multi-step reasoning prompts and rubrics for Frontier AI Labs.

• Review work submitted by experts, provide feedback, and approve it for client delivery
• Document edge cases and assist in improving learning materials
• Create complex multi-step reasoning prompts & rubrics for Frontier AI Labs

AI Trainer at Labelbox (Alignerr) (2025-04 – Present)

Stress-test AI models using different prompting techniques. Identify and document logical fallacies and detect model failures. Build rubrics for AI model search queries and perform rubric-based evaluation of AI responses.

• Stress-test AI models using different prompting techniques such as Chain of Thought and Tree of Thought
• Identify and document logical fallacies, detect model failures, and flag potentially unsafe and biased responses to support policy alignment and safety improvements
• Build rubrics for AI model search queries (router model), and perform Rubric-based evaluation of AI Responses

Senior Quality Auditor (QA) at Appen (2024-04 – Present)

Oversee the quality of 100+ contributors and auditors across dozens of markets. Perform individual spot checks, manual reviews, and detect market-wise error trends. Provide manual feedback, maintain documentation, and build learning resources.

• Oversee the quality of 100+ contributors and auditors across dozens of markets by performing individual spot checks, manual reviews, and detecting market-wise error trends
• Provide manual feedback, maintain documentation, and build learning resources to address error trends detected through root cause analysis
• Collaborate with Quality Managers to take proactive steps to mitigate fraud, ensure compliance, and support continuous quality improvement initiatives

Project Lead at Upwork Inc (2024-12 – 2025-02)

Helped design and refine a domain rating project workflow for the world's largest search engine company. Trained and onboarded 9 new talents across 6 countries. Implemented continuous quality monitoring and created self-serve learning materials.

• Helped design and refine a domain rating project workflow for the world's largest search engine company
• Trained and onboarded 9 new talents across 6 countries on the process workflow
• Implemented continuous quality monitoring through 1-on-1 feedback, spot-checks, RCA, and created self-serve learning materials ensuring guideline compliance
• Calibrated on quality KPI improvement and process optimization with the client

AI Content Evaluator at Centific (2024-04 – 2024-08)

Crafted single and multi-turn AI prompts and evaluated AI responses on various criteria. Performed SxS evaluation of AI model responses on overall quality using the Likert Scale.

• Crafted single & multi-turn AI Prompts and Evaluated AI Responses on Instruction Following, Relevance, Completeness, Accuracy, Safety, Context Awareness, Writing Style & Tone
• Performed SxS evaluation of the AI model Responses on overall quality using the Likert Scale

Quality Analyst (QA) at Appen (2023-08 – 2024-04)

Managed quality of over 600 expert talent networks, including the entire roster from India in English and Hindi. Performed Root Cause Analysis, answered rater tickets, conducted Trust & Safety investigations, and built learning resources.

• Managed quality of over 600 expert talent networks, including the entire roster from India (English & Hindi)
• Performed Root Cause Analysis (RCA), answered Rater tickets using Zendesk, did Trust & Safety Investigation, built learning resources, & provided target feedback for continuous quality improvement

Search Quality Rater at Appen (2016-06 – 2023-08)

Evaluated search results and AI responses for various quality metrics. Identified and flagged offensive and unsafe content across multiple formats. Awarded International Rater Spotlight for consistent top-tier quality scores.

• Evaluated search results and AI responses for Needs Met, Relevance, Truthfulness, Safety, Helpfulness, and Accuracy
• Identified and flagged offensive/unsafe content across images, videos, and texts by enforcing the given policy documents
• Awarded International Rater Spotlight for consistent top-tier quality scores

BA (English) - IGNOU (2025 - 2028)

Completed 4 years of coursework in B.Tech in Mechanical Engineering – Lovely Professional University (2010-08 – 2014-05)

10+2 in Science – Bijni Bandhab HS School (2008-08 – 2010-05)