Sreya J.

About Sreya J.

AI/ML Engineer with 4+ years of expertise in architecting and deploying production-grade intelligent systems. Specialized in Generative AI and Agentic RAG frameworks (LangGraph, LlamaIndex), consistently delivering a boost in retrieval accuracy and reduction in hallucinations for enterprise knowledge bases. • Proven track record of building high-throughput ETL/ELT pipelines using Databricks (Delta Lake) and Snowflake (Snowpark) to process multi-terabyte datasets (2TB+). Expert in optimizing data movement and feature engineering workflows that reduce inference latency by 25.• Advanced proficiency in A/B testing and statistical hypothesis testing to validate model performance. Leveraged Power Analysis and Bayesian inference to drive a 15% increase in marketing ROI and 32. • Experienced in orchestrating the complete ML lifecycle on AWS (SageMaker, Lambda) and Azure ML. Expert at implementing automated retraining loops, CI/CD pipelines (GitHub Actions), and containerization (Docker) to serve thousands of real-time API requests daily with reliability. • Accomplished in fine-tuning state-of-the-art models (RoBERTa, LLaMA 3, Claude 3.5) using PEFT/QLoRA for specialized tasks in finance and healthcare. Additionally, engineered computer vision pipelines (Mask R-CNN) that increased object detection precision in industrial automation. • Skilled in translating complex AI outputs into high-level business strategy through interactive Tableau/Power BI dashboards and stakeholder presentations. Focused on reducing reporting turnaround by 30% and providing clear, actionable insights to cross-functional teams.

Experience

1. Ai Engineer

   DXC Technology

   Sep 2024 — Present · US

   Architected and deployed production-grade Agentic RAG pipelines using LangGraph and LlamaIndex, leveraging Databricks for distributed data processing of 2TB+ unstructured datasets, which improved retrieval accuracy.• Engineered automated ELT pipelines within Snowflake using Snowpark, migrating legacy Python scripts into stored procedures to reduce data movement latency and enable real-time feature serving for LLM agents.• Developed a multi-step document intelligence system using RoBERTa and NER, automating the extraction of structured insights from complex financial records with precision, resulting in a 35% increase in document throughput.• Orchestrated MLOps workflows on AWS (SageMaker, Lambda) to automate model versioning and A/B testing deployments; implemented SHAP analysis to ensure model interpretability and compliance with enterprise data governance.• Spearheaded the transition to a Lakehouse architecture on Databricks, utilizing Delta Lake for ACID transactions and data versioning, which cut reporting turnaround time for cross-functional stakeholders.• Designed and executed systematic A/B tests to evaluate prompt engineering strategies (Few-shot vs. Chain-of-Thought), achieving a lift in response relevance and ensuring 98% alignment with domain-specific knowledge bases.

Education

• University of New Haven

  Master of Science, Computer Science

• Chaitanya Bharathi Institute Of Technology

  Bachelor of Engineering, Information Technology