FlowCV Logo
Professional Experience

Klass Engineering and Solutions

Software Engineer (AI Engineering)
08/2023 – present
  • •Designed system and implemented an app to showcase LLM Orchestration Capabilities. This opened opportunities forfuture work with AI Agents delegating and executing tasks
  • •Modularizing UI into a reusable base Chatbot framework to reduce technical debt and streamline development, cuttingUI development and integration time for future teams.
  • •Architected and developed an in-house centralized model weights caching, to address scaling and MLOps challenges.This resulted in 3% (2TB/70TB) of disk space savings and 5 days of time savings per developer
  • •Analyzed third party codebase to identify and debug critical RAM and VRAM leak, resulting in 100% improvement ofefficiency
  • •Implement database caching of chunking and vectorization stages of RAG for production environments, improvingefficiency by at least 100%
    • Projects
    02/2025 – 03/2025
  • •Built an AI-powered system that automatically handles multi-part queries, significantly reducing manual effort,improving response times, and increasing operational efficiency in customer interactions
  • •Implemented automated error handling and conflict resolution, ensuring more reliable booking processes, minimizingdisruptions, and enhancing overall user experience
  • •Introducted LLM-based testing and Langsmith tracing to ensure high-quality, consistent outputs from AI agents,significantly reducing troubleshooting time and improving overall system stability and performance
    • 09/2024 – 11/2024
  • •Evaluated trade-offs between NoSQL, Milvus, PostgreSQL and PGVector extension, accessing the additionalcomplexity required to implement cross database consistency
  • •Evaluated trade-offs between using AWS Lambda + Amplify and EC2 for deployment, opting for EC2 to simplify local andserver testing for both backend and UI
  • •Leveraged IaC with Terraform to automate provisioning of AWS EC2, ECR, and security policies, enabling cost-effectiveand reproducible setup and teardown of cloud resources
  • •Optimized document ingestion and vectorization on a small EC2 instance, using a rolling window approach to avoidmemory constraints and preserve context between chunks
    • Skills
    Cloud & Devops: GCP, Grafana, Prometheus, GH Actions|Databases: Cassandra, Redis, Sqlite|Languages & Frameworks: Python, Fastapi, Langgraph, Langchain, Typescript, React, CPP