p
prashantdhya516

Percy

@prashantdhya516

I turn your AI idea into a working system, fast

India
Inglese, Hindi
Alcune informazioni sono riportate in lingua inglese.
Chi sono
6 years building AI systems - not side projects, actual production pipelines running at enterprise scale. I've built call center classification systems that process thousands of transcripts daily, RAG pipelines for document intelligence, LLM automation workflows, and GPT-4 integrations for US telecom clients. I don't just write code - I figure out what the business actually needs, then build something that holds up. If you have a messy problem that you think AI can solve, bring it to me. I'll tell you straight if it's viable, and build it if it is.... Continua a leggere

Competenze

p
prashantdhya516
Percy
offline • 

Portfolio

Esperienza lavorativa

Microsoft

Senior Data Scientist

Microsoft • Full time

Nov 2022 - Present • 3 yrs 6 mos

Embedded with a Fortune 500 telecom client in the US, building and deploying production-grade AI systems end-to-end. Led development of an LLM-powered call classification engine — processing thousands of transcripts daily across multi-level category taxonomies using GPT-4 and advanced prompt engineering. Built speech analytics pipelines on ASR output data, extracting structured business signals from unstructured call recordings at scale. Designed SQL-based validation frameworks for product data integrity, and delivered executive-ready reporting systems used by senior business stakeholders. Work spans the full stack — data ingestion, LLM integration, pipeline orchestration, and stakeholder output.

American_Express

Data Scientist

American Express • Full time

Jul 2019 - Oct 2022 • 3 yrs 3 mos

Worked as a Data Scientist at a telecom-focused analytics firm in Oman, building ML systems to reduce churn and improve customer retention across a large subscriber base. Developed end-to-end churn prediction models using supervised learning - handling feature engineering, model selection, validation, and deployment into production. Built customer segmentation and recommendation systems that drove personalized engagement at scale. Collaborated closely with business and product teams to translate raw telecom data into actionable intelligence, delivering models that directly influenced retention strategy and revenue outcomes.