Shreyas
Revankar
I build systems that make work disappear. From LLM screening pipelines that slash 70% of manual effort, to production platforms serving 500+ users — my work is defined by outcomes you can measure.
Engineering systems that think faster than humans can act
My edge is simple: I obsess over the gap between what software does and what it should eliminate. When I built an AI screening pipeline at ASU, I didn't stop at "it works" — I drove it to under 30 seconds per candidate with structured JSON outputs a recruiter can act on immediately.
My stack spans Python, Java, Go, React, and SvelteKit, but what drives me is the layer beneath — LLM prompt engineering, retrieval-augmented generation, and workflow orchestration with tools like n8n and Gemini AI. I care about measurable outcomes, not impressive-sounding architecture.
I'm pursuing my Master's at Arizona State University while actively seeking SWE internships or co-ops in AI integration, backend infrastructure, or full-stack product development.
Systems with measurable impact
Every project below has a number attached to it. That's intentional.
AI Resume Screening Pipeline
From inbox to shortlist in under 30 seconds
Orchestrated an end-to-end AI hiring pipeline using n8n and Gemini AI. The system automatically pulls resumes from Gmail and Google Drive, performs semantic evaluation against job descriptions, and delivers structured JSON outputs — match scores, skill deltas, and recruiter summaries — without a single manual step.
CropSync — Smart Farming Platform
Predictive AI + blockchain for agricultural intelligence
Full-stack microservices platform combining AI-driven crop forecasting with blockchain-backed data integrity. Integrated real-time weather and market APIs with a multi-layer cache that eliminated 60% of redundant external calls. Containerized with Docker and deployed via CI/CD to AWS, cutting release errors by 35%.
Agile Scrum Simulation Engine
Modeling real sprint dynamics across 200+ team scenarios
A Java-based backend simulation engine using state-transition logic to model Agile task lifecycles. Built role-differentiated Swing UI for 6 Agile personas with enforced permission scopes. Integrated full CI/CD via GitHub Actions, achieving 40% improvement in simulation execution consistency across all sprint scenarios.
Production-grade under real load
Full Stack Developer Intern
InternshipShipped production frontend and backend features for a platform serving 500+ active users, delivering measurable latency improvements and security hardening under real concurrent load.
- →Architected a SvelteKit frontend with a Flask REST backend for a multi-workflow platform, reducing user-facing configuration errors significantly through improved state management and UI clarity.
- →Engineered JWT-based authentication with session invalidation and token expiry policies, eliminating unauthorized access attempts across all production environments with 500+ active sessions.
- →Redesigned backend data models with targeted C++ optimizations, cutting average API latency 25% — from 400ms to 300ms — under high concurrent production load.
- →Designed a role-based access control layer across 4 user tiers, enforcing least-privilege permissions at the API layer and reducing the platform attack surface.
- →Translated product briefs into detailed API contracts and component specs, enabling parallel development across 2 engineers and shipping 3 features ahead of the 10-week deadline.
- →Automated UI regression testing via Tcl scripts, reducing regression test cycle time and increasing release confidence across all delivery iterations.
M.S. Software Engineering @ Arizona State University — Actively seeking SWE internships 2026
Built for depth, shipped for scale
Hover a category to explore skill depth. Levels reflect real project application.
AI & Automation
Languages
Frameworks & Tools
Cloud & DevOps
Let's build something that matters
I'm actively seeking SWE internships or co-ops for 2026, especially roles involving AI integration, backend infrastructure, or full-stack product development. If your team is building meaningful systems, I'd like to talk.