Hari

Hi, I am

Harihara
Vardhan K

Google Summer of Code '26 @ SugarLabs
100x Developer powered by AI Agents.

Download CV
Scroll to explore

About me

At the intersection of code and curiosity

I build things somewhere between AI, systems, and late-night curiosity. Most days I am working on agents, backend tools, or web apps; the rest of the time I am probably skating through Bangalore traffic pretending it is less chaotic than my codebase. I like tech because it feels a lot like skateboarding: you fall a lot, learn balance slowly, and every once in a while something clicks so well it feels effortless.

AI

Agents, LLM apps, automation

Systems

Backend, APIs, performance

Tech Stack

(hint: hover over a key)

Experience

My professional journey.

01

Google Summer of Code '26

Sugar Labs · Music Blocks · Remote

May 2026 — Sept 2026

Building a Git-based backend for Music Blocks so thousands of creative coding projects can move from a legacy MySQL store into durable, searchable GitHub-backed repositories with ownership, history, and room to grow.

  • Designed the migration path for 10,000+ user projects from MySQL to GitHub with persistent version history and ownership.
  • Built a streaming migration pipeline that processes projects in under 5 seconds each with in-memory transformation and resumable checkpointing.
  • Engineered a SQLite + REST API layer that brings search and retrieval latency from roughly 200ms via GitHub API reads to under 1ms locally.
  • Designed around GitHub's 5,000 req/hr API limits by offloading read-heavy browsing and search to SQLite across 10K+ repositories.
  • Implemented SHA-256 deduplication to eliminate redundant storage and improve migration efficiency.
  • Developed 7+ production-grade endpoints for search, browse, like, and publish flows while preserving legacy feature parity.
GitGitHub APISQLiteREST APIsPythonSHA-256Backend
Visit

02

CPP Intern · Quantum Computing

Hewlett Packard Enterprise (HPE) · Remote

Feb 2026 — Present

Working on quantum-assisted scheduling for heterogeneous memory systems, using optimization models to decide where tasks should live across DRAM and CXL-like memory tiers.

  • Designed a quantum-assisted scheduling engine with RQAOA to optimize task placement across DRAM and CXL memory tiers.
  • Formulated scheduling as a QUBO problem, modeling memory latency from 80-400ns alongside capacity constraints for task allocation.
  • Built a hybrid optimization system with Qiskit, OpenQAOA, and classical solvers for recursive combinatorial problem reduction.
  • Simulated CXL-like memory architectures using NUMA and controlled memory binding with numactl for realistic performance evaluation.
  • Developed an end-to-end optimization to scheduling to execution pipeline for heterogeneous memory conditions.
  • Evaluated latency, throughput, and task completion time across multi-tier memory system configurations.
C++Quantum ComputingRQAOAQUBOQiskitOpenQAOANUMACXL

01 · project

Curia – AI

A full-stack meeting summarizer reducing manual documentation by 60%.

ReactExpressLLMFlask

02 · project

MakeMore – AI

An AI name generator using PyTorch and Neural Networks for character-level language modeling.

Neural NetworksNLPPyTorch

03 · project

Lucy – AI

An interactive AI chatbot platform with AI-generated avatars, collaborative sketching, and location-based exploration.

ReactGemini AIFlask

04 · project

Chat Web App

A scalable backend for a real-time chat service using Flask and Socket.IO, reducing message latency by over 80%.

FlaskSocket.IOJavaScript

contact

Let's talk?

If what you've seen interests you, the keyboard is ready for the first message.

Email copied · hariharavardhan1234@gmail.com
Open mailtoGitHubLinkedIn