University of Illinois Urbana-Champaign
Master of Science in Computer Science
- GPA: 3.96 / 4.00
- Relevant coursework: Cloud Computing Applications, Deep Learning for Computer Vision, Software Engineering
Aryaman Nasare
Software Engineer building
reliable systems that scale
I'm Aryaman Nasare — CS Grad @ UIUC (GPA 3.96/4.00). I build backend services, automation, and applied AI features with an emphasis on reliability, performance, and clean engineering.
Backend & APIs
Automation / CI
AI + RAG
Open for roles across US
About Me
Who I am and what I do.
I'm a computer science graduate student at the University of Illinois Urbana-Champaign, focused on building intelligent software that is reliable, scalable, and production-ready. My background spans software engineering, automation, data systems, and AI.
I started my journey in India, where I worked on engineering projects and gained hands-on experience in cloud systems and automation before moving to the U.S. to deepen my foundation in algorithms, systems, and applied AI.
I'm currently seeking software engineering roles where I can work on meaningful systems, learn from strong engineers, and ship work that matters at scale.
Education
Academic background and coursework highlights.
Savitribai Phule Pune University
Bachelor of Computer Engineering (Honors in Data Science)
- GPA: 8.74 / 10.00
Experience
Work across automation, data systems, and backend engineering.
May 2025 – Dec 2025
Champaign, Illinois
Graduate Teaching Assistant
University of Illinois Urbana-Champaign
- Developed data systems assignments covering SQL schema design and indexing, NoSQL/graph systems (MongoDB, Neo4j), and Docker-based data curation workflows for 550+ students across two graduate courses.
- Engineered a Bash automation script using the GitHub REST API to parse user CSVs and batch-provision organization access, saving 15+ hours of manual onboarding effort.
May 2025 – Aug 2025
Remote (Miami, Florida)
Software Engineer Intern (Technology Track)
RIA Advisory
- Built 60+ end-to-end automation flows with Selenium to validate a SaaS test automation platform, using dynamic waits and strict assertions to reduce flakiness and expand regression coverage.
- Engineered a LangChain RAG pipeline indexing 200+ pages of internal technical guides into a FAISS vector store, enabling 20+ interns to self-resolve environment issues and reducing troubleshooting time by 33%.
Jan 2024 – Jul 2024
Navi Mumbai, India
Software Engineer Trainee
Reliance Jio Platforms Ltd.
- Orchestrated regression strategies for the JioPartnerWorld application (1M+ users), validating cloud-hosted microservices via Jenkins CI/CD pipelines and Jira-tracked end-to-end scenarios, reducing production defect leakage by 15%.
- Built a Selenium automation suite executed via AWS CodeBuild with parallelized cloud execution to cut test execution time by ~90% (saving 6+ minutes per cycle).
- Led root cause analysis across distributed systems by analyzing AWS CloudWatch logs and SAP IDoc failures, improving MTTR by 40% through cross-team collaboration.
Featured Projects
Code that solves real problems.
Full-Stack
InsightHub
AI-powered team collaboration with secure docs + RAG chat/search.
Designed and built a full-stack, multi-tenant platform for secure document management, team activity notifications, and AI-assisted chat. Implemented JWT auth with centralized Express middleware for team-membership/creator authorization, plus CORS allowlisting for strict tenant isolation. Added an embeddings + vector similarity RAG pipeline over uploaded documents for context-aware answers inside private workspaces.
Deep Learning
GhibliDream
Few-shot image generation via DreamBooth fine-tuning (Stable Diffusion 2.0).
Built an end-to-end few-shot fine-tuning pipeline enabling Studio-Ghibli-style generation from only 5–7 image–text pairs; containerized with Docker for reproducible reruns. Automated LLM-assisted caption rewriting across an 800+ image dataset and used a two-token identifier strategy to improve subject/background disentanglement, reaching 0.914 foreground CLIP similarity.
AI Agents
SWE-Agent
Autonomous code repair system using tool-driven LLM workflows.
Extended the Mini-SWE-Agent framework to resolve GitHub issues by building 10 structured tools (regex search/replace, linter, memory search) for code navigation and patching via a tool-call protocol. Achieved a 70% resolution rate (28/40) on SWE-bench Verified (Django, Matplotlib), improving 2.3× over baseline (12/40) with iterative validation inside a Docker sandbox.
Technical Arsenal
Technologies and tools I work with.
Programming
Backend / Cloud
AI / ML
Automation & Testing
Let's build something that works at scale
Want to collaborate, chat about a role, or discuss a project? Reach out to me here...