Venkatesh Shivandi
Profile
Venkatesh Shivandi
Portland, ME | 207-252-5059 | venkateshshivandi18@gmail.com
LinkedIn • GitHub
Applied AI Engineer and Full-Stack Developer with a focus on building production-ready AI systems and rapid MVPs. Experienced in architecting multi-agent workflows, optimizing LLM costs, and delivering high-impact software that bridges the gap between raw data and actionable clinical or business insights.
Education

M.S., Applied Machine Intelligence

Northeastern University — Portland, ME | Dec 2025

Coursework: AI Systems, Responsible AI, MLOps, AI Visualization, Enternprise Analytics

B.S., Computer Science

Osmania University — Hyderabad, India | June 2021

Coursework: Data Structures & Algorithms, Web Development

Professional Experience

AI Developer Intern — Aosenuma

Apr 2025 – Sep 2025 | Portland, ME

  • Architected and deployed an AI-powered finance automation platform (React/FastAPI/PostgreSQL) processing 1,000+ monthly receipts with 99%+ uptime.
  • Built RESTful services orchestrating Azure Document Intelligence and GPT-4, improving extraction accuracy from 73% to 89%+ while reducing costs by 60% via prompt caching.
  • Developed a production-ready Chrome extension with JWT authentication, reducing manual data entry time by 70% for vendor portals.
  • Engineered agent-assisted browser automation using Playwright and WebSockets for autonomous claim submissions with human-in-the-loop oversight.

Junior Software Engineer — Cognizant

Apr 2022 – Jul 2023 | Hyderabad, India

  • Migrated legacy SAS analytics to Azure Synapse/ML, contributing to a $5M annual infrastructure cost reduction.
  • Built Azure Data Factory pipelines for mainframe/DB2 data ingestion, delivering clean datasets for ML model training and batch scoring.
  • Monitored propensity model accuracy and drift metrics in Azure ML, validating outputs to ensure consistent marketing campaign performance.
  • Optimized SQL queries, reducing data preparation turnaround time and improving pipeline reliability.

Key Projects

DocForge AI (Multi-Agent Documentation System)

Python, FastAPI, CrewAI, RESTful APIs (Published on PyPI)

  • Published a production-grade Python package with 500+ downloads, featuring a CLI and RESTful API for CI/CD integration.
  • Architected agent coordination using CrewAI for parallel processing and error recovery, automating multi-format technical documentation.

Waynew (AI Learning Platform)

Python, FastAPI, React, TypeScript, PostgreSQL, Claude API

  • Built a full-stack learning platform using Claude API to generate personalized curricula and interactive paths with real-time streaming.
  • Delivered sub-30s end-to-end course generation while supporting concurrent user sessions through scalable REST APIs.

Roux-MMC (Clinical Risk Prediction)

Machine Learning, Python, SHAP

  • Developed ML models outperforming the STS benchmark in predicting eight post-operative cardiac surgery outcomes (e.g., 0.911 AUROC for prolonged ventilation).
  • Created clinician-friendly visualizations using SHAP for risk factor breakdowns, bridging ML outputs with surgical decision-making.

Certifications
Oracle Certified: Generative AI Professional
Microsoft Certified: Azure Data Engineer Associate (DP‑203)
Microsoft Certified: Azure AI Fundamentals (AI‑900)
Technical Skills
Programming & Frameworks: Python, TypeScript, JavaScript, FastAPI, React, SQLAlchemy, Pydantic
AI & LLM Systems: RAG, CrewAI, LangChain, OpenAI/Claude APIs, Google Vertex AI, LoRA/PEFT fine-tuning, MCP
Backend & Databases: RESTful APIs, PostgreSQL, MongoDB, Neo4j, Redis, SQL Optimization
Cloud & DevOps: Azure (ML Studio, Synapse, Document Intelligence), AWS, Docker, Kubernetes, CI/CD
© 2026 Venkatesh Shivandi