Experience

Analyzed sustainability disclosures from 500+ retail companies, applying ESG taxonomy to classify and benchmark firms across three tiers; produced structured datasets and visual summaries used in 2 quarterly reports.

Identified and resolved 100+ scraper-ETL issues (malformed URLs, 404s, HTML parsing errors) via regex-based validation - improving data reliability by 25% and stabilizing uptime.

Automated a preprocessing layer to normalize tabular inputs and flag inconsistencies, cutting preprocessing time by 20% and manual review by 30%.

Built multimodal ML workflows integrating speech transcriptions, product images, and price data for automated MSME credit scoring; improved ETL throughput by 40%.

Fine-tuned Whisper ASR on regional datasets, achieving 31.5% lower Word Error Rate; engineered PyTorch object detection capable of processing 10K+ images for financial document classification.

Designed Tableau dashboards surfacing transaction trends and pricing anomalies; implemented SQL validation logic ensuring consistency between source and processed data.

Performed backend and API testing for cloud deployments across three client production systems, collaborating in an agile cross-functional team.

Built Selenium automation scripts and executed 200+ API tests via BrowserStack; resolved 15+ vulnerabilities across responsiveness, latency, and multi-OS security vectors.

Improved system response time by 30% (5s → 3.5s) by redesigning the UI testing workflow and refining request handling logic.