AI/ML Platform & Cloud Engineering Lead

Sai Raghav Keesara

Ó+1-814-***-**** R *************@*****.*** /raghavkeesara /raghavkeesara

Technical Skills

Languages Java, Python, C, C++, TypeScript, Kotlin, Ruby, Scala, HTML Technologies Spring, REST, Docker Containers, Kubernetes, Flask, Angular, Spark, PyTorch, TensorFlow, MLOps Databases Elasticsearch, MongoDB, SQL, MySQL, PostgreSQL, MSSQL AWS Step Functions, Lambda, ECS, AWS Glue, EC2, CloudWatch, Experience

Salesforce San Francisco, CA

Lead Member of Techical Sta Oct 2025 - Present

[Cloud Economics & Capacity Management] Leading engineering e orts for Horizon Agent, a Slack-based LLM platform for cost and capacity insights, driving 80+ monthly active users, 1000+ conversations, and 80% e cacy.

De ning and delivering Horizon Reporting Service, Salesforce's second generally available internal AI product, enabling LLM-powered exploratory data analysis and automated insight delivery for decision makers.

Driving expansion of the internal Horizon Data Platform transformation layer into a product-oriented, automation- rst data platform, enabling sub-1-day turnaround for new data products by non-engineers.

Providing technical leadership in AI product development, platform architecture, and agile execution across CECM, partnering with cross-functional teams to align roadmaps, quality, and adoption.

Amazon Web Services Seattle, WA

Software Development Engineer II May 2021 - Sep 2025

[AWS AI/ML Rekognition & Bedrock] Migrated Existing ML Infrastructure hosting Computer Vision Deep Learning models at scale to Nvidia Triton inference Server to achieve 33% better throughput for same latency

Build an Inference work ow that classi es documents and ingests into Index that is used for by LLMs to answer questions on Amazon Q for project Bedrock Data Automation Link

Built an Auto-scaling pipeline that reduces ECS Fleet size by predicting the load thus reducing operating costs by 30%

[Amazon Search Rufus] Built Amazon's conversational AI Shopping experience widgets on the Search Results Page to be fetched within 150ms latency budget

Implemented UX to refetch LLM streaming content on animation or streaming interruption for Rufus Desktop Implementation

[Amazon Ads] Enhanced the creation rate of advertiser ad campaigns on Amazon.com by 20% through the implementation of creative recommendations, resulting in an additional $200 million in revenue. Link

Processed legacy Ad creative data to assess acceptance rates, extract valuable insights, and e ciently synchronize the information into new data storage systems, ensuring seamless data migration for downstream services

CiteSeerX @ Pennsylvania State University University Park, PA Graduate Research Fellow August 2019 - May 2021

Led the design and development of project scalable next generation CiteSeerX, the world's rst academic digital library.

Designed and implemented new Elasticsearch schema as well as Migrated existing data from SQL

Achieved 4x higher Ingestion throughput by developing a novel Data Extraction and Ingestion pipeline

Contributed to several open-source repositories in SeerLabs

VMware Bangalore, India

Software Engineer Jan 2018 - Aug 2019

[VMware Workspace One] Designed and developed the back-end architecture allowing users to rate applications, customize features and portal branding for enterprise Identity and MDM software .

Achieved near 100% up-time by implementing a fall-back JWT cookie based user session management mechanism to address DynamoDB and Memcached connection failures

Improvised release process by eliminating DevOps' dependency by automating the turning on of SaaS features to deployment phase using principles of Infrastructure as Code

Amazon Bangalore, India

Software Engineer Jul 2017 - Dec 2017

[Native Shopping Ads] Developed a Full Stack Web app that renders key-phrase based sponsored ads for comparative shopping websites

Extracted curated ad impression details from Amazon S3 logs using Scala/Spark and checked sponsored ad coverage

Implemented a Python deep web-crawler and computed product-centricity scores based on ad publisher's domain Education

Pennsylvania State University University Park, PA Master of Science in Computer Science; GPA: 3.79/4.0 Aug. 2019 { May, 2021

Operating Systems, Algorithms Design & Analysis, Distributed Systems, Cloud Computing, Deep Learning

Birla Institute of Technology and Science (BITS) Pilani, India B.E (Hons.) in Computer Science & M.Sc (Hons.) in Mathematics; GPA: 8.35/10.0 Aug. 2013 { June. 2017 Publications

"COVIDSeer: Extending the CORD-19 Dataset", 20th ACM Symposium on Document Engineering 2020, San Jose, CA, USA

"Building a Scalable, and Sustainable Service for Scholarly Big Data", 2021 IEEE International Conference on Big Data

Contact this candidate