Srinivas Pendyala

Led 10+ customers, resulting in $5+ million deal sizes, and published an APN prescriptive guide, driving customer adoption and revenue growth.

Some of the top customers I've worked with include Henry Schein, IBM Multicloud Management Platform (MCMP), JPMorgan, Citigroup, HDFC Bank, Visa, Mastercard, FedEx, Palo Alto Networks, Telesign, Scentsy, Cordial, T-Mobile, Coinbase, Ekata, Nexmo (now Vonage), Blue Cross Blue Shield Association, and Onclusive.

I drove strategic discussions on solution architecture based on economies, including total cost of ownership (TCO) discussions for optimal ROI on cloud investments. I also tested and certified the Amazon EKS integration with Redis Enterprise's custom Kubernetes operator and automated the AWS infrastructure setup using CloudFormation templates and Terraform scripts.

Integrated Amazon Bedrock with Redis as a vector database. I collaborated with AWS engineering teams to integrate Redis as a knowledge base for RAG, providing Redis best practices and API support and establishing CI/CD infrastructure.

• Guided AWS teams on the Jedis API to invoke vector similarity searches and on KNN vs. ANN searching algorithms.
• Provided technical guidance and direction on vector indexing methods like flat and Hierarchical Navigable Small World (HNSW).
• Helped make design choices for vector distance metrics, such as cosine vs. L2 vs. inner product (IP), to measure vector similarity.
• Utilized Amazon Titan as the embedding model and Anthropic Claude 3 as the LLM model.

Implemented a RAG pipeline and built a chatbot application using LangChain and sentence transformers for financial analyst use cases.

I implemented an object storage-based data lake. The knowledge base consisted of financial data such as SEC filings, trade rules documents, trading compliance policy PDF documents, and Word documents. Next, I built an MLOps pipeline that sniffs on the data lake to sync data to the vector database periodically, such as every few hours and every night. Finally, I implemented LLM embedding to vectorize the data into the Redis Enterprise cloud vector database and LLM semantic caching to save on LLM computation cycles.

Built a RAG application to assist healthcare professionals in their diagnosis and improve patient outcomes.

With a knowledge base of medical literature, patient records, medical research papers, and medical treatment and compliance documents, I implemented an LLM semantic caching benchmark that showcased 1,600 to 4,000 times faster latency speeds, saving on LLM computation cycles.

Designed the solution architecture of an AI/ML-based fraud detection solution on AWS for customers and published best practices as an APN blog and reference architecture. I also trained data using XGBoost and RCF ML algorithms.

Additionally, I used Feast with Redis as a feature store to develop ML operations. I also displayed single-digit latencies with Amazon SageMaker, leveraging Redis as a feature store for fraud detection in finance services use cases.

Built a data lake based on Cloudera's distribution of Hadoop (CDP) for multiple customers.

• Enabled raw data ELT into the data lake from streaming, transactional, and batch data sources into Hadoop using Sqoop and Kafka.
• Modeled data for operational databases using HBase on Hadoop.
• Built ETL pipelines using data wrangling and Trifacta preparation features to reduce redundant analytical cycles.
• Designed the solution and architected end-to-end near-real-time streaming and batch pipelines to veto Cloudera Hadoop on the Informatica, Talend, and Pentaho data engineering stack.
• Enabled data analytics using Tableau and MicroStrategy by running ad hoc analytical queries using Apache Impala.
• Developed time series data modeling for IoT use cases on GE Historian.

Developed and deployed Hadoop security automation tooling in-house using Ansible, Bash, Apache Sentry, Hive, Kerberos, TLS, and encryption at rest with AWS Key Management Service (KMS). I also built Jenkins CI/CD pipelines using Docker containers for end-to-end Data warehousing experience.

Developed Altus SDK for customers to leverage Cloudera Altus in the AWS and Azure public cloud. I worked on the API design for SDK and simulated Amazon's Elastic MapReduce (EMR) API design for portability.

Worked on an API monetization program, defining the enterprise architecture to enable data discovery across the enterprise for Walmart and Sam's Club via RESTful APIs with big data Hadoop datasets.

I used Apache Camel for polling and Oozie workflow to invoke Hive ETL actions to prepare and enrich data. I also used EL methodologies and architected the enterprise data hub on Hadoop for early analytics. Lastly, I designed and built data pipelines to offload ETL workloads from the traditional data warehouse into Hadoop.

Designed the solution architecture and developed a targeted personalization engine for Sam's Club subscribers.

I used enterprise data sets for data mining and built and deployed a campaign management system using an enterprise data layer accessible over APIs. I used Cassandra and HBase for operational data and Greenplum for metadata needs.