Arjun Kantamsetty
Software Engineer
Passionate about AI, software engineering, and the intersection of technology and creativity.
Projects
A collection of my work in software engineering and AI.
Upright
Oct. 2025
Built a posture-monitoring desktop app that collects anonymized session data to deliver posture insights and fine-tune pose estimation models, enabling personalized ergonomics feedback for 100+ users.
OpsCopilot
Aug. 2025
Built OpsCopilot, a multi-agent AI system that integrates with incident management and observability tools to help on-call engineers detect, diagnose, and resolve production issues faster.
TrueCaption
Feb. 2025
Built an end-to-end platform that fine-tuned an ASR model (60% → 95% accuracy) to transcribe STEM lectures with domain-specific terminology, addressing hidden accessibility barriers in education. Presented to an audience of 500+.
Experience
My professional journey in software engineering and AI.
AI Infrastructure Engineer
WEX • Portland, ME
May 2025 – Present
- Enabled distributed training for WEX’s modeling teams by creating autoscaling EKS nodepools with Karpenter, configuring Ray on AKS, and integrating SageMaker IAM to orchestrate CPU/GPU workloads across Kubernetes pods in multi-cloud environments.
- Improved claims adjudication accuracy from 70% → 85% on date-of-service extraction within two weeks of joining the project, reducing manual review requirements by 2.25M claims annually.
- Implemented an end-to-end observability stack with Helm, Kubernetes, and ArgoCD, delivering visibility across all environments for services supporting millions of customers — executed 3× faster than comparable organization initiatives.
- Led several company-wide trainings on building AI applications with RAG and multi-agent workflows, equipping both technical and non-technical teams with the skills to securely integrate AI into internal workflows; trained over 200 engineers across multiple departments.
AI Engineering Intern
BPRHub • San Francisco, CA
Feb. 2025 - Jul. 2025
- Built Octo, a production-grade agentic AI system for a manufacturing compliance platform, accelerating manufacturing policy audits and boosting compliance through integrated evaluation and optimization pipelines by 45%
- Achieved a 25x speedup in RAG inference by semantic chunking, optimized batching, and embedding model evaluation, and benchmarking vector database performance for low-latency retrieval.
Software Engineering Intern
Levo.ai • Austin, TX
Aug. 2024 – Jan. 2025
- Built an LLM pipeline to enhance API documentation readability by 30%, reducing onboarding time for client developers and cybersecurity teams.
Get In Touch
Interested in collaborating or have a question? Feel free to reach out.