Rohan Waghmare

Software Engineer

Palo Alto, California, United States5 mos experience

Key Highlights

Achieved 10x cost reduction in AI pipelines.
Reduced workflows from hours to minutes for thousands of engineers.
Engineered scalable backends for AI/ML systems.

Stackforce AI infers this person is a Backend-focused Software Engineer with expertise in AI/ML and Data Engineering.

Contact

Skills

Core Skills

Machine LearningSoftware DevelopmentData EngineeringFull-stack DevelopmentBackend Development

Other Skills

PythonDeep LearningGCPAWSSAP BTPHuggingFaceArgoCDKServeFastAPISPARQLMCPSAP HANALangGraphGPT-4PKCE

About

Software Development Engineer building scalable backends and AI/ML-powered systems using Python. Worked across fine-tuned LLMs, RAG pipelines, Knowledge Graphs, and agentic workflows on GCP, AWS, and SAP BTP. Notable achievements: cut costs by 10x, reduced workflows from hours to minutes, and built tools used by thousands of engineers. I am in it for the tech, problem solving, and the memes. 📧 Email: waghmarerohan30@gmail.com 🌐 Portfolio: rohanwaghmare.com 🔗 GitHub: github.com/ron103 👾 Leetcode: leetcode.com/u/rohanwaghmare/

Experience

5 mos

Total Experience

5 mos

Average Tenure

5 mos

Current Experience

Sap

Software Developer

Jan 2026 – Present · 5 mos · Palo Alto, California, United States · On-site

Replaced a multi-hop GPT-4 pipeline at 10x lower cost by fine-tuning DeepSeek (QLoRA) on SAP BTP, achieving 100% domain extraction accuracy at 8s inference via HuggingFace, ArgoCD, KServe, and FastAPI.
Cut developer query-to-documentation time by 92% (1hr → 5min) for 5,000+ engineers by building a RAG pipeline over a SPARQL Knowledge Graph with confidence-scored retrieval, exposed as an MCP server (SSE, JSON-RPC 2.0) for Claude/Cline on SAP HANA.
Engineered a multi-step sequential agentic pipeline (LangGraph + GPT-4) on SAP BTP with SPARQL Knowledge Graph retrieval, backend update tool access via MCP, and enterprise-grade auth/auditing through PKCE + SAP IDP.

PythonDeep LearningMachine LearningGCPAWSSAP BTP+8

Michigan health information network shared services (mihin)

Software Engineer

Mar 2025 – Jan 2026 · 10 mos · Lansing, Michigan, United States · Remote

Engineered a data pipeline in Python processing 500K+ records across 10+ organizations using AWS ECS Fargate, S3, EventBridge, DynamoDB, with Docker/Terraform deployment and cross-month aggregation logic.
Reduced processing latency 40% by refactoring S3 Select API with conditional ScanRange and multipart parallelization via ThreadPoolExecutor, implementing 20MB chunking across multi-GB files.
Built real-time SQS monitoring with Timestream, CloudWatch, and Grafana, identifying redundant Lambda executions and idle queues, reducing AWS costs 1.8%.
Built a Python + Selenium scraper to migrate NPI directory data at scale, eliminating 60% of manual processing.
Identified and solved security vulnerabilities in AWS Cognito flows using DevTools and ThunderClient.

PythonAWSDockerTerraformDynamoDBEventBridge+6

Binghamton tech collective

Software Engineer

Aug 2024 – Mar 2025 · 7 mos · Binghamton, New York, United States · On-site

Built Shop Bing, a marketplace web app with JavaScript (React.js) and Firebase, enabling secure student-to-student transactions; onboarded 100+ users in the first two weeks.
Implemented trust and safety features (.edu verification, block/unblock, reporting), reducing fake sign-ups by 95% and 200+ item listings by semester’s end; coordinated Agile sprints, peer reviews, and CI/CD with GitHub Actions.

JavaScriptReact.jsFirebaseAgileGitHub ActionsFull-Stack Development

Flow

Backend Engineer

Jul 2024 – Aug 2024 · 1 mo · Wilmington, Delaware, United States · Remote

Designed smart ingestion filters (job title, tenure, company size/funding) that reduced low-value contacts by 35% and improved lead relevance by 53%, directly enhancing an LLM/RAG-powered sales assistant workflow.
Built and optimized scalable Python(Django)/PostgreSQL pipelines, integrated APIs from LinkedIn, PitchBook, and Crunchbase, and containerized backend services with Docker.

PythonDjangoPostgreSQLDockerBackend Development