Suryateja Chalapati

Built scalable data pipelines, AI automation, and deployed production LLM agents for Data Platforms and Supply Chain Analytics.

• Built and deployed a fully autonomous data ingestion agent on GCP handling batch formats (CSV, fixed-width, TXT, GitHub exports) end-to-end — raises its own PRs, monitors Tekton pipeline runs, tracks BigQuery load status. Reduced ingestion setup from months to minutes.
• Deployed a ServiceNow incident resolution agent that autonomously triages alerts, enriches context, and delivers actionable recommendations to engineers via Microsoft Teams with an embedded live chat interface.
• Designed and implemented LLM and multimodal pipelines on GCP (Vertex AI, Hugging Face) for entity recognition, summarization, and document classification at scale.
• Developed scalable ingestion and migration workflows moving ~20 TB of data into BigQuery via Dataflow, Airflow, and GCS.
• Built ETL pipelines using Airflow, Python, and SQL to ingest batch and streaming data from APIs, Teradata, and Oracle.
• Applied GenAI models for NLP tasks using Transfer Learning and Deep Learning on GCP.

Developed natural language models, pipelines and metrics using unstructured data for multiple high-profile clients.

• Built spaCy Transformer NLP models achieving 93% accuracy with custom entity extraction and semantic similarity components.
• Delivered precision/recall metrics directly enabling $500K in new business for a major client.
• Fine-tuned ALBERT Transformer models using AWS SageMaker and HuggingFace across 7 TB of de-identified transcript data.
• Designed end-to-end MLOps pipeline improving model operational efficiency by 60%.

• Estimated Kickstarter page consistency at 7% using Cosine Similarity and Semantic Analysis on unstructured text fetched via API.
• Identified that 36% of Reddit posts in 2020 were health-related via web scraping and NLP analysis.
• Built ETL pipelines on GCP with Airflow, migrated on-prem data to cloud data warehouse; Topic Modeling (LDA, TFIDF) yielded highest correlation of 38%.

Optimized medical clinic operations by building ML models and metrics.

• Built a Random Forest model on patient data improving clinic operating efficiency by 38%.
• Extracted data from S3 and Redshift; created multiple databases in AWS Glue Catalog using Glue Crawlers.
• Automated model training in AWS SageMaker and built Tableau dashboards for operational insights.

• Reduced freight delay risk by 15.3% via automation tools and risk assessment metrics.
• Predicted labor productivity with 82% precision (ALPS) leading to EU freight optimization.
• Increased annual shipments by 34% over 6 months via A/B Testing and process development.
• Deployed 3PL logistics sites with 92% success rate in collaboration with supply chain teams.

• Optimized data imputation in S3, reducing costs by $90K and rollout times by 15.3% using AWS Athena.
• Built ETL jobs on AWS Glue to load vendor data from multiple sources with cleaning and transformation.
• Worked across AWS services: S3, EC2, Glue, Athena, Redshift, EMR, Kinesis, DMS, SNS, SQS.