Aryan K funinkina

23, a Delhi kid who moved to Bangalore and somehow fell for the weather and the community. These days I’m building company brain and an agentic OS at Libra. Off the clock I contribute to open source and Linux, ricing my desktop and building software for humans.

Resume
Aryan K

Experience

5 roles in two years, mostly voice and LLM backends. Newest first.

Jul 2026Present
Libra AI
SWE 1
HSR Layout, Bangalore
  • Shipping core surfaces of Libra’s agentic workspace (TypeScript, React): the Tasks agent UI with live run activity, plan-step navigation and inline doc/slide previews, generative UI in chat, an AI slide deck generator with 20+ themes, and an onboarding revamp. Migrated the app from React Router to TanStack Router.
  • Built the desktop app’s meeting stack: live captions, a meeting transcription bot, in-meeting notes and offline recording; evaluated streaming STT providers and added voice-fingerprint speaker separation.
  • Built and hardened 15+ OAuth integrations (Microsoft 365, Salesforce, HubSpot, Stripe, Jira, Confluence, GitHub, Linear, Figma, Workday, Google Workspace), fixing token lifecycles and short-lived tokens; shipped the M365 connector for an enterprise POC.
  • Overhauled the AI email assistant and daily digest, scaled the document ingestion pipeline (file parser + self-hosted llama.cpp distiller), and own the marketing site’s SEO/AEO on Cloudflare, plus a Chrome extension.
Apr 2026Jul 2026
Zyka AI
Backend Development Intern
Remote
  • Built and designed an audio/image annotation platform (Next.js, React) for annotators to tag large-scale multimedia datasets, architecting the UX/UI for high-throughput tagging workflows.
  • Developed an API aggregator backend unifying multiple LLM inference providers, implementing credit-based usage metering, request routing, and rate limiting for cost-tracked inference access.
Jun 2025Apr 2026
Superdash AI
Backend Engineer Intern
Indiranagar, Bangalore
  • Optimized and deployed edge ML models for voice interruption detection, audio de-noising, and forced word alignment, plus self-hosted TTS (5× speedup, 1000ms → sub-200ms); served via WebAssembly, vLLM, and ONNX Runtime, sustaining sub-100ms inference across concurrent live calls.
  • Engineered scalable voice agent backend systems in Go and Python, owning full lifecycle from API design through production deployment; refactored WebSocket layer to cut voice call latency by 80% (500ms → 100ms).
  • Designed and shipped an event-driven automation pipeline (akin to n8n) as a core product feature, fully owning the initiative from client co-design and architecture through testing and production rollout; directly drove new client acquisition and revenue growth.
Dec 2024Feb 2025
Weya AI
Python Developer Intern
Noida, Uttar Pradesh
  • Built a real-time voice agent backend using FastAPI and WebSockets, scaling to 100+ concurrent connections with stable p99 latency under load.
  • Profiled 10+ system metrics across LLM inference and speech pipelines to track down bottlenecks and improved streaming responsiveness and reduced end-to-end latency meaningfully.
  • Added integrations for Calendar, Meet, WhatsApp, and Twilio. Also built a microservice to convert S3 documents into embeddings.
May 2024Jul 2024
NIT Jalandhar
Summer Research Intern
Jalandhar, Punjab
  • Designed, trained, and evaluated neural network models to predict waterborne diseases from 20+ real-time water quality parameters that hit ~95% predictive accuracy through systematic hyperparameter tuning and feature engineering.
  • Co-authored a peer-reviewed research paper on the model architecture, training pipeline, and experimental results, also got accepted and published in an indexed journal.

Highlights

390+ stars and 2K+ users across GitHub. Three worth calling out.

Linux-native webcam effects engine

OpenEffects

A GTK4 app over a headless daemon with a PipeWire/GStreamer pipeline. ONNX Runtime inference dispatches to CUDA, ROCm or Vulkan for sub-50ms portrait blur, center stage, studio light and background replacement. Shipped as a Flatpak on Flathub.

Rust80 stars2K+ users
CLI secrets manager

DeadEnv

Replaces plaintext .env files with the OS keychain on macOS, Linux and Windows behind biometric auth. Exports are AES-256-GCM + Argon2id with error collapsing against oracle attacks, plus a git audit log.

Go16 stars
Native CUPS driver for a Windows-only printer

Ricoh SP 200 driver

Captured and decoded the printer's USB protocol from the Windows driver, then wrote a CUPS filter with JBIG1 (ITU-T T.82) compression that reconstructs the full job lifecycle and prints multi-page documents.

C
Everything else on GitHub

Writing

Linux, systems and whatever I'm building, read by 1000+ people a month.

All 14 posts

Projects

Tools I built because I wanted them, mostly for Linux desktops and terminals.

Image

Deadenv

Replacing .env files with OS-native secret storage

gosecurityclidevtools
Image

Weekly Commits

GNOME Extension to see your weekly GitHub commits in top bar

linuxgnomejavascriptextension
Image

RSync Backup Script

A customizable shell script to backup home folder to extrnal drive

shell scriptinglinux
Image

Query MD

Ask and query your markdown notes using AI 🤖

llmgenAItui
Image

Bloop

Get fun and quirky insights from your whatsapp chats

nextjsgolang
Image

Spectacle Ocr Screenshot

A simple Qt application that integrates KDE Spectacle screenshot tool with Tesseract OCR to extract text from screenshots as well QR codes.

tesseractqtlinux
All 9 projects

Commits

Public GitHub activity over the last year. Work and private repos not included.

contributions by @funinkina

Beyond work

Subscribe via RSS