What happens when you give someone not just tools but the freedom to imagine systems that think, decide, and act on their own? On May 29–30, 2026, that question came to life inside F1Soft’s ARL, which transformed into a high-energy innovation space for the F1 ARL Hackathon 2026: Agentic AI Edition. Across the F1Soft, engineers, innovators, problem-solvers, and AI enthusiasts gathered with a shared curiosity and ambition: to explore what software could become when intelligence is no longer reactive, but autonomous.
Set in the theme “Building Autonomous Agentic AI Systems (100% Local)”, the two-day hackathon pushed participants beyond conventional automation. Teams were challenged to design intelligent agents capable of planning, reasoning, and executing tasks independently, systems that could not just assist workflows, but manage them end-to-end.
Problems to Prototypes
The seeds of the Hackathon were sown months ahead, with an internal campaign called “What’s Your Problem?”. Employees across the organization were encouraged to share real problems from their daily work, bringing forward inefficiencies, bottlenecks, and unmet needs. What began as scattered conversations gradually evolved into a structured innovation pipeline, where problems were not just acknowledged but reframed as opportunities for transformation.
From these real-world problem statements, the idea for the hackathon was developed as a platform to turn challenges into innovative solutions. A total of 25 teams from across different companies within the group registered to participate, spanning domains such as engineering, operations, quality assurance, ERP, DevOps, and product development. Following a selection process, 11 teams comprising nearly 50 participants advanced to the final round.
Over two intensive days, these teams set out to turn ideas into reality designing, building, and demonstrating autonomous AI agents running entirely on local infrastructure. The outcomes demonstrated not only strong technical innovation but also the potential of enterprise AI solutions that prioritize privacy, operational independence, and intelligent automation.
Competing Teams and Project Showcase
The hackathon featured a diverse set of teams building innovative autonomous agent solutions across multiple domains:
HireMinds: Autonomous CV Review Agent
HireMinds, developed by Asna Shakya, Bikal Shrestha, Hemant Basnet, Pukar Shrestha, and Shubhekshya Adhikari, is a multi-agent recruitment assistant designed to streamline and enhance the hiring process. The solution automatically screens resumes, removes personally identifiable information (PII) to help reduce bias, evaluates candidates against job requirements, ranks applicants, and generates interview preparation kits for recruiters.
Aegis: Autonomous DevOps Agent
Aegis: Autonomous DevOps Agemt, developed by David Dangol and Bikram Sapkota, is an on-premise DevOps automation solution designed to enhance system reliability and operational efficiency. The agent is capable of analyzing system logs, identifying issues, suggesting remediation steps, and generating automation scripts with human approval safeguards for critical operations.
Memoria: Autonomous Personal Knowledge Agent
Memoria, developed by Sarun Thapa, Saugat Neupane, Grishma KC, Saurab Kharel, and Simran Panday, is a secure offline “second brain” designed for managing and reasoning over personal and organizational knowledge. The system enables users to store, search, and retrieve information locally using vector databases, while leveraging ReAct-based reasoning to generate context-aware insights.
Meeting Memo: Agentic AI Meeting Notes
Meeting Memo, developed by Nitesh Ghimire, Prabin Chaudhary, Bijay Shrestha, and Adarsh Ghimire, is a secure local-first system designed to transform meeting transcripts into structured and actionable insights. The agent converts meeting transcripts into structured summaries, action items, and decision logs while ensuring all processing remains local and secure.
FONEAI: Autonomous Software Engineering & Security Agent
FONEAI, an autonomous software engineering & security agent developed by Mikal Shrestha, Aakash Ghimire, and Prajwal Acharya is an intelligent engineering assistant designed to support software development and security workflows. The system is capable of reviewing codebases, generating pull requests, performing compliance checks, and assisting with security assessments, all while maintaining human oversight for critical decisions.
ERP Copilot
ERP Copilot, developed by Pushkar Kumar Sah, Rajib Sigdel, Anuj Poudel, Nitisha Bhattarai, and Anil Ghale, is a natural language interface designed to simplify interaction with enterprise resource planning systems. The solution enables users to execute complex ERP workflows, generate reports, and query enterprise data using conversational inputs, while maintaining strict tenant-level security controls.
KubeSentinel AI
KubeSentinel AI, developed by Rajib Dahal, Krisha Silwal, and Suben Pandey, is a Kubernetes incident-response agent designed to improve the reliability and observability of cloud infrastructure systems. The solution automatically analyzes cluster telemetry to detect anomalies, identifies potential root causes of infrastructure failures, and helps reduce troubleshooting time through local reasoning capabilities.
V.I.S.T.A: Autonomous Multi-Modal QA Agent
V.I.S.T.A, Autonomous Multi-Modal QA Agent, developed by Sushil Dhakal, Sanil Manandhar, Ayur Adhikari, and Prashan Dristi, is a zero-code testing platform designed to simplify and automate software quality assurance processes. The system enables users to perform API testing, UI testing, performance evaluation, and security checks using natural language commands, eliminating the need for manual scripting.
CodeMind: Autonomous Local Code Intelligence CLI
CodeMind, developed by Rashmita Subedi, Nilima Maka, Sunaina Ghimire, Shyam Bista, and Nischal Phuyal, is an AI-powered terminal-based assistant designed to help developers understand and manage large codebases efficiently. The system enables users to identify bugs, analyze complex project structures, and perform safe refactoring directly from the command line, all while ensuring that code remains local and is not transmitted to external systems.
Travel Sathi
Travel Sathi, developed by Anubhuti Shah, Bipin Neupane, Yural Pokhrel, Bikram K.C., and Gigyasha Niroula, is an autonomous travel planning system that leverages multiple coordinated agents to create personalized travel experiences. The platform intelligently designs end-to-end itineraries by optimizing routes, estimating trip costs, and tailoring plans based on user preferences.
Lavish Mystery: Agentic AI Voice Support Platform
Lavish Mystery, Agentic AI Voice Support Platform, developed by Sandesh Shrestha, Rupesh Sunuwar, and Bibek Gurung, is a voice-enabled autonomous customer support system designed to streamline and enhance service operations. The platform utilizes specialized agents to handle different aspects of support, including query routing, troubleshooting, escalation management, and analytics.
Winners of the Hackathon
Winners of the Hackathon were announced after a rigorous evaluation process based on innovation, technical execution, business relevance, and autonomous capabilities. The top honor was awarded to Team Crazy AI, Travel Sathi, which secured first place along with a prize of NPR 2,00,000 for its advanced multi-agent travel planning system. The First Runner-Up was Team ReAct Rangers, V.I.S.T.A, receiving NPR 1,00,000 for its zero-code, multi-modal QA testing platform. The Second Runner-Up was Team Himalaya, Memoria, awarded NPR 50,000 for its secure offline personal knowledge agent. Collectively, these winning projects showcased strong technical depth, practical real-world applicability, and the growing potential of Agentic AI in transforming enterprise workflows.
Beyond a Competition: A Platform for Innovation
The F1 ARL Hackathon 2026 was more than a competition, it was a learning ecosystem, an innovation catalyst, and a preview of the future of work.
Participants explored a wide range of autonomous agent applications, including reasoning systems, orchestration engines, DevOps automation, quality assurance frameworks, recruitment assistants, travel planners, and enterprise copilots.
As Agentic AI continues to reshape industries globally, the hackathon reaffirmed a key insight: innovation thrives when individuals are empowered to experiment, collaborate, and build meaningful solutions. Many of the prototypes developed during the event have the potential to evolve into impactful enterprise products that address real business challenges.
