Rajitha V.

Key Skills

Software

Don't disclose

Data Annotation Tech

Labelbox

Scale AI

Dataloop

Top Subject Matter

Global payments operations training and operator performance modeling

Transaction verification quality automation and early ML risk scoring

NLP-based payment query classification and operational intelligence

Top Data Types

Text

Document

Image

Top Task Types

Classification

Entity (NER) Classification

RLHF

Text Generation

Question Answering

Text Summarization

Evaluation/Rating

Computer Programming/Coding

Data Collection

Object Detection

Fine-tuning

Freelancer Overview

AI/ML Training Engineer – Global Payments Operations. Brings 21+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose, Internal, and Proprietary Tooling. Education includes Master of Science, New Science College, Kakatiya University and Bachelor of Science, Vaagdevi College, Kakatiya University. AI-training focus includes data types such as Text, Computer Code, and Programming and labeling workflows including Classification and Entity (NER) Classification.

Labeling Experience

Senior AI/ML Engineer – Intelligent Operations & Automation (HSBC Global Service Centre)

TextEntity (NER) Classification

Built and deployed NLP-powered systems to resolve payment queries and classify user intents, supporting intelligent operational decisioning. Developed predictive and anomaly-detection models that convert operational text and case information into actionable outputs for downstream teams under regulatory constraints. Implemented production MLOps pipelines for monitoring, governance, and retraining to sustain model performance. • Engineered multi-class query resolution using Random Forest/XGBoost across 1,200+ investigations monthly, reducing false positives by 35%. • Created NLP-driven root-cause analysis automation using clustering and NLP to eliminate 250+ inbound queries per month. • Designed BERT-embedding-based workflow/intent classification with 94% production accuracy across 25+ payment scenarios. • Implemented model governance and monitoring dashboards for drift, accuracy, latency, and business KPIs using MLflow and CI/CD practices.

2006 - Present

AI/ML Training Engineer – Global Payments Operations

Don't discloseTextClassification

Designed adaptive learning and predictive analytics for payment operations to support data-driven skill development for operators. Built and applied ML proficiency and decisioning systems that forecast performance from historical resolution data to enable early intervention and consistent training outcomes. Authored training guidance to standardize AI-assisted workflows and decision support across teams. • Developed collaborative-filtering-based learning curriculum for 14 payment workflows and 50+ global operators. • Built logistic regression proficiency prediction achieving 95% accuracy for 6-week forecasting of at-risk operators. • Developed decision tree systems from historical resolution data to increase operator self-resolution by 40% and reduce time-to-proficiency. • Automated quality scoring to reduce manual assessment time by 60% and improve consistency across training cohorts.

2012 - 2015

Payment Systems Engineer – Verification & Quality Automation

Don't discloseClassification

Transitioned verification work into automated quality assessment and early risk-scoring intelligence to support training and operational compliance. Engineered rule-based transaction verification logic and prototyped ML approaches for transaction risk and fraud pattern detection. Used anomaly/outlier style heuristics to reduce submission errors and improve downstream rework metrics. • Built rule-based verification engine processing 3,000+ transactions monthly with complex validation logic and zero compliance breaches. • Prototyped logistic regression transaction risk scoring (entered production later) and supported departmental AI adoption. • Developed statistical outlier detection patterns for real-time irregular transaction behavior and error-prediction heuristics. • Improved operational consistency by reducing operator submission errors by 18% and downstream rework by 20% YoY.

2009 - 2012

Education

Work History