Back to Developers
SOUNDARYA S

SOUNDARYA S

AI Data Analyst

Chennai, India
85
Profile Score

About

AI Data Analyst with 2+ years at Amazon specialising in LLM response evaluation, large-scale dataset quality assurance, and annotation pipeline optimisation. Consistently delivered 98%+ accuracy on high-volume labelling tasks across text, image, audio, and video modalities. Hands-on contributor to model improvement cycles — from surfacing hallucinations and policy violations to rewriting model outputs and refining annotation guidelines that raised team-wide quality standards. Bringing a rare combination of data rigour, AI domain fluency, and QA leadership to a data analyst role.

Skills & Expertise (18)

Factuality Verification Intermediate
8.0/10
2
Years Exp
Hallucination Detection Intermediate
8.0/10
2
Years Exp
Policy Violation Flagging Intermediate
8.0/10
2
Years Exp
Response Ranking Intermediate
8.0/10
2
Years Exp
SenSai Intermediate
8.0/10
2
Years Exp
Microsoft Azure SLA Management Structured Feedback Loops Pattern Recognition Dataset Validation Guideline Refinement Edge Case Documentation Accuracy Auditing Peer review Amazon SageMaker Ground Truth V7 Cvat Labelbox

Work Experience

AI Data Analyst (Data Associate)

Amazon

Feb 2024 - Feb 2026

Maintained 98%+ quality accuracy across high-volume, multi-modal annotation tasks (text, image, audio, video), consistently meeting or exceeding Amazon's data quality benchmarks on production AI/ML datasets. Reduced LLM hallucinations and policy violations by evaluating, ranking, and rewriting model responses during RLHF-style feedback cycles — directly improving the quality of training signals fed back into the model pipeline. Improved team-wide annotation consistency by systematically identifying edge cases and ambiguous scenarios, documenting findings with reproducible examples, and collaborating with QA leads to update guidelines — reducing annotation disagreements across the team. Lowered team rework rates through structured peer reviews and quality audits, delivering actionable written feedback that enabled colleagues to self-correct errors before submission, cutting QA rejection cycles. Sustained 100% SLA adherence across concurrent annotation projects by managing personal productivity targets without compromising data integrity — supporting timely delivery of training data to downstream ML model teams. Operated independently on complex edge cases using the SenSai annotation tool, interpreting ambiguous guidelines without escalation and contributing documented decisions that fed back into updated team-wide protocols.

Education

Bachelor of Computer Applications (BCA)

2020 - 2023 · Afghanistan

Certifications

Microsoft Azure Cloud Computing — Internship Certification

· 2023

Interested in this developer?

Profile Score Breakdown

📷 Photo 10/10
📄 Resume 10/10
💼 Job Title 10/10
✍️ Bio 10/10
🛠️ Skills 20/20
🎓 Education 10/10
⏱️ Experience 5/15
💰 Rate 0/5
🏆 Certs 5/5
Verified 5/5
Total Score 85/100

Profile Overview

Member sinceMar 2026

Availability Details

Visa Status

Citizen

Relocation

Open to Relocation

Skills (18)

Factuality Verification Hallucination Detection Policy Violation Flagging Response Ranking SenSai Microsoft Azure SLA Management Structured Feedback Loops Pattern Recognition Dataset Validation +8 more