ArnavAngarkar
GSoC '26 @ FOSSASIA · Speaker · 4x Hackathon Winner
Full-stack engineer and ML researcher. I ship production open source, build autonomous systems, and deploy models that replace million-dollar solutions.
About Me
Arnav Angarkar— CS undergrad at IIIT Dharwad. GSoC '26 at FOSSASIA. ISRO IROC AIR 3. Linux Foundation Scholar. 4x hackathon winner. Published researcher in quantum error correction.
Speaker at FOSSASIA Summit 2026. I build full-stack platforms, deploy ML models that replace commercial solutions, and architect real-time streaming infrastructure.
Education
Indian Institute of Information Technology, Dharwad
B.Tech in Computer Science and Engineering
- Focus areas: Machine Learning, Systems Programming, Algorithms.
Indian Institute of Information Technology, Dharwad
Minor in Quantum Information Science & Technology
- Quantum Computing minor covering quantum algorithms, error correction, and quantum information theory.
Loyola High School & Junior College, Pune
SSC
- PASCH Juku Scholarship Alumni (2021 batch) — awarded for German language excellence.
- Member of PASCH – Schulen: Partner der Zukunft, an international network of schools.
Work Experience
May 2026 – Presentvoxbento.com
Google Summer of Code 2026 – FOSSASIA
RemoteGSoC Contributor | VoxBento
- Selected as a GSoC 2026 contributor to architect VoxBento from scratch – a distributed, low-latency multilingual audio platform for conference-scale virtual and hybrid events, deployed and tested on Linux server infrastructure.
- Designed and implemented a concurrent multi-channel audio routing system using WebRTC, Jitsi, and MediaMTX – core multimedia/audio-codec pipeline work; engineered synchronization primitives to coordinate real-time audio streams across parallel language channels without drift or race conditions.
- Built end-to-end horizontally scalable speech pipelines: concurrent Speech-to-Text, translation, subtitle generation, and Text-to-Speech; coordinated via asynchronous event-driven processing with Redis pub/sub for inter-service synchronization.
- Engineered scalable backend microservices with Python, FastAPI, PostgreSQL, and Redis on Linux; designed distributed interpreter management and language-routing logic for conference-scale multilingual audio delivery.
- Presented distributed systems architecture and engineering trade-offs as a speaker at FOSSASIA Summit 2026, and live-demoed VoxBento at Wikimania Paris 2026 to attendees from the global Wikimedia community.
October 2025 – Presentgithub.com/fossasia/eventyay
FOSSASIA
RemoteProject Maintainer | Eventyay
- Maintainer of Eventyay, a 6k+ GitHub star open-source conference management platform used by organizations including Wikimedia, PyCon, and numerous international conferences.
- Showcased Eventyay in person at Wikimania Paris 2026 to multiple conference organizers and community members.
- Led development and maintenance across Django and Jinja codebases on Linux, reviewing contributions, merging pull requests, maintaining production stability, and delivering features spanning ticketing, scheduling, and organizer workflows.
- Invited to FOSSASIA Summit 2026 (Bangkok) to present open-source contribution workflows and platform engineering practices.
June 2025 – July 2025
Poonawalla Fincorp
Pune, On-siteML/AI Intern
- Deployed an ML-powered identity verification system integrated into fraud prevention workflows, replacing a commercial solution valued at ~Rs. 2 Cr/year.
- Built custom self-hosted data infrastructure for high-speed vector operations and streamlined identity verification.
- Optimized algorithmic performance for decision accuracy and inference efficiency at enterprise scale.
June 2025 – Present
QWorld
RemoteQuantum Intern | Project Lead
- Leading a QWorld project (PAQ) to build an interactive, game-based learning platform for quantum computing.
- Designing modules on Bell states, CHSH inequality, quantum tunneling, decoherence, and quantum foundations.
- Managing a cross-functional team of 5 to integrate storytelling, quantum logic, and game mechanics.
February 2025 – December 2025campusdabba0.netlify.app
CampusDabba
DharwadFounder & CEO
- Founded and scaled a pre-incubated campus marketplace startup for students to buy, sell, and exchange goods and services.
- Built the full platform end-to-end — Node.js, TypeScript, Next.js, MongoDB, WebSockets — from first commit to production deployment.
- Managed business operations, user acquisition, and product roadmap while leading technical architecture decisions.
- Won 1st Place at DevHack '25 (IIT Dharwad) competing against 500+ applicants across all years.
June 2025 – December 2025
IISc (Indian Institute of Science) – RIMM Lab
BangaloreML/AI Intern | Smart Mirror Project
- Engineered a hybrid virtual try-on pipeline integrating multiple deep learning models for full-body outfit previews with high photorealism.
- Built real-time CV pipelines with MediaPipe and body landmark detection for live webcam-based garment overlay.
- Explored Neural Anthropometry for 3D body rendering and adaptive garment fitting for personalized digital fashion.
- Designed modular ML APIs and GPU-based inference setups for scalable, high-performance try-on experiences.
November 2024 – Aug 2025
ISRO Grand Challenge – All India Rank 3
IndiaProject Manager | Command & Control Systems Engineer
- Led a team in a 10-month deep-tech initiative to design and deploy a fully autonomous UAV Drone—selected among Top 3 teams from 2000+ nationwide.
- Built a custom MAVLink-based GCS with real-time telemetry dashboards, state-aware mission mode transitions, and 3D flight visualizations using WebSockets and Three.js.
- Integrated edge-AI hardware (Jetson Nano), Pixhawk, LiDAR, and depth cameras for SLAM-based navigation, dynamic path planning, and autonomous obstacle avoidance.
- Developed computer vision pipelines using YOLOv5, Depth Anything, DINOv2, and PyTorch for safe landing zone detection and multi-modal perception.
December 2024 – March 2025
IISc – CPDMed TBI (Aarnaq Project)
BangaloreML/AI Intern
- Fine-tuned YOLOv8 for Automatic Number Plate Recognition on a GoK-DULT funded urban mobility project.
- Built reproducible training pipelines with data versioning, augmentation strategies, and hyperparameter sweeps.
- Optimized inference for edge deployment — ONNX export, FP16 quantization, TensorRT acceleration.
November 2024 – April 2025
Brahmand Exploration
PuneResearch Intern & AI Developer
- Built and deployed an AI-powered conversational bot using NLP for automated customer query resolution on the company website.
- Developed predictive models for internal workflows and contributed to front-end/back-end development with performance optimization.
- Revamped brahmandexploration.com into a dynamic digital encyclopedia with improved UX, content accessibility, and site performance.
2022 – Present
Bug Bounty Programs
RemoteSecurity Researcher
- Active contributor to the Netflix, Cursor, 1Password, Udemy, and Google Bug Bounty programs, successfully identifying and reporting security vulnerabilities.
- Discovered and responsibly disclosed vulnerabilities across web applications, APIs, and authentication flows, earning acknowledgements from security teams.
- Developed deep expertise in OWASP Top 10 attack vectors including XSS, IDOR, CSRF, and authentication bypass techniques.
Achievements
PresenterWikimania Paris 2026
Live-demoed VoxBento and Eventyay to global Wikimedia community members and conference organizers.
GSoC ContributorGoogle Summer of Code 2026
GSoC contributor with FOSSASIA; one of ~1,200 selected globally.
View Project →SpeakerFOSSASIA Summit 2026
Invited speaker at one of Asia's largest open-source tech conferences. Presented on architecting real-time multilingual streaming infrastructure.
AIR 3ISRO IROC Grand Challenge
All India Rank 3 in ISRO's national autonomous UAV challenge. Led team through 10-month initiative competing against 2000+ teams.
1st PlaceHACKMAZE '26 — Telemetry Power Governor RTL
Telemetry-driven Power Governor RTL IP for dynamic power management in SoCs. Hardware FSM for adaptive power state transitions.
View Project →
1st PlaceDevHack '25 — CampusDabba
Won first place at DevHack at IIT Dharwad with CampusDabba, competing against 500+ applicants across all years.
View Project →
3rd PlaceHAL Aerothon — RagVault
Fully offline, privacy-focused RAG system for secure enterprise knowledge management. Zero external API dependencies.
View Project →
FinalistHack2Future '25
Only freshman team to qualify for the final round out of 400+ applicants. Received special mention.
Projects
VoxBento – Distributed Multilingual Streaming Infrastructure
May 2026 – Present🌐 voxbento.comWebRTCJitsiMediaMTX+5 more
- Architected VoxBento from scratch as a GSoC 2026 contributor – a distributed, low-latency multilingual audio platform for conference-scale virtual and hybrid events.
- Designed and implemented a concurrent multi-channel audio routing system using WebRTC, Jitsi, and MediaMTX to coordinate real-time audio streams across parallel language channels without drift.
CerebrumAI – Offline-First Multimodal Healthcare Triage Engine
January 2025 – PresentFlutterReactTypeScript+10 more
- Built a privacy-first, AI-powered triage system enabling early detection of mental and cognitive disorders via multimodal inputs (text, image, behavior, voice).
- Designed modular architecture with quantized LLMs (Phi-2, Bio-Medical LLaMA-3-8B) and singleton pattern for real-time model switching and performance optimization.
CampusDabba
January 2025🌐 campusdabba0.netlify.appNode.jsTypeScriptExpress.js+5 more
- Developed a dynamic campus marketplace for students to buy, sell, and exchange goods/services within the college community.
- Implemented secure user authentication and real-time chat functionality using WebSockets.
IIIT Dharwad Official Website
2025 – Present🌐 iiitdwd.ac.inNext.jsTypeScriptAngular+4 more
- Member of the official IIIT Dharwad Website Team — contributing to development and maintenance of the institute's public-facing website.
- Building and improving pages across academics, campus life, and information sections using modern front-end and back-end frameworks.
RagVault – Offline AI-Powered RAG Knowledge Portal
January 2026PythonFastAPIReact+10 more
- Built a 100% offline, enterprise-grade RAG system delivering GPT-4 level document intelligence with zero data leaving the network — solving the privacy-vs-AI dilemma for 78% of enterprises blocked by data regulations.
- Engineered a hybrid retrieval pipeline combining BM25 keyword search (0.3×) with FAISS semantic vector search (0.7×) plus BGE-Reranker-v2-M3 neural reranking, achieving 87% Precision@5 (+19% over vector-only).
Customer Service Conversation Analyzer
March 2024PythonNLPMachine Learning+4 more
- Developed a sophisticated real-time conversation analysis system for customer service calls.
- Implemented speech-to-text using OpenAI's Whisper and sentiment analysis with BERTweet.
AutoML-MLOps Platform Prototype
October 2024H2O.aiMLflowStreamlit+4 more
- Designed an end-to-end AutoML platform with automated model training, experiment tracking, and task orchestration for machine learning workflows.
- Implemented a user-friendly interface using Streamlit and integrated LLM-based bot features for user guidance.
BMP Image Filter
August 2024CCommand LineMakefile
- Developed a command-line tool that applies various filters (grayscale, sepia, reflection, blur) to BMP image files.
- Implemented algorithms to process image data efficiently, ensuring high-quality output while optimizing for speed.
Dynamic API for NoSQL Query
December 2024MongoDBPyMongoPostman+3 more
- Developed a backend API that performs upsert operations (single and bulk operations) in MongoDB. The API allows users to specify collection name, filter conditions, and update fields.
- Implemented flexible input handling for unstructured data in filter and update operations.
Binary Classification Web App
December 2024StreamlitPythonScikit-learn+2 more
- Built a powerful web application using Streamlit to predict binary outcomes, allowing users to classify data points as belonging to one of two classes.
- The app leverages multiple machine learning classifiers, including Support Vector Machine (SVM), Logistic Regression, and Random Forest, to deliver accurate predictions on any given dataset.
Health Hub 360
September 2024FlaskJinjaHTML+3 more
- Developed a digital platform to assist patients in managing booking appointments, and connecting with healthcare providers.
- Implemented features for automated inventory management of medicines in hospitals and clinics.