On to the 2nd Phase ! DevTrails 2026

Panthadeep Majumder — Thu, 02 Apr 2026 10:59:48 +0000

Dev Blog: The "QuickShield" Journey – From Theoretical Planning to a Hard-Earned Model

🛡️ The Vision: Protecting those who Power the City

Our project, QuickShield, was born out of a simple but urgent necessity: the gig economy is the backbone of urban India, yet its workers are the most exposed to environmental and social volatility. We set out to build an AI-driven Dynamic Underwriting Engine that doesn't just offer insurance, but actively calculates risk in real-time. By analyzing localized climate data and urban disruption patterns, QuickShield enables a "Protective Premium" model—scaling safety costs precisely when the environment becomes adversarial.

🧗 The Phase 2 Grind: The Reality of "App-Ready"

Getting selected for Phase 2 was an incredible milestone, but it was also the moment the "honeymoon phase" of planning ended. We quickly realized that a conceptual model on a slide deck is worlds apart from a production-ready system.

The transition from Plan to Model has been our steepest climb yet. We moved away from simple "if-else" logic to a robust XGBoost Predictive Engine. We had to solve the "Inference Gap"—building a technical bridge so our Node.js backend could communicate with our Python ML brain seamlessly. Every step forward revealed a new layer of complexity, from handling non-serializable JSON data to ensuring our city encoders were perfectly synced across different environments. It’s been a journey of moving from "what if" to "how exactly."

📝 The Quiz: A Moment of Truth

Yesterday, we faced a project-related quiz that felt like the ultimate pressure cooker. We were locked in, focused, and precise—answering every single technical and conceptual question with 100% accuracy. In those moments, there was a brilliant, shimmering semblance of hope that we had secured a spot in the Top 3.

When the rankings were revealed, however, we found ourselves "far off" from the leaders. It was a sobering reminder of the caliber of competition we are up against. But while the leaderboard didn't reflect our position where we wanted it, our perfect score on the questions gave us something more valuable: Validation. It proved that our technical foundation is rock solid. We know the "Why" and the "How"—now we just have to execute the "What."

🚀 The Path Ahead: Fuelled by Optimism

Despite the setbacks and the grueling technical hurdles, we aren't slowing down. We’ve successfully moved our project from a folder of CSVs to a functional, integrated ML system. The heatmap we generated doesn't just show data; it shows a model that understands the pulse of Indian cities—the monsoons in Mumbai, the AQI spikes in Delhi, and the chaos of urban festivals.

We are choosing to be optimistic. We have a working "Brain," a solid "Bridge," and a team that gets the answers right when it counts. Whether we hit the Top 3 or not, QuickShield is becoming a reality, and we are ready to build the future of worker safety, one line of code at a time.

BuildInPublic #QuickShield #InsurTech #MachineLearning #FounderJourney #GigEconomy #Python #NodeJS

Gig-Workers having trouble? No issue Insurance is here!

Panthadeep Majumder — Wed, 18 Mar 2026 17:53:24 +0000

Building AI-Powered Insurance at Guidewire DevTrails 2026

1. Our Journey Begins

We are excited and confident to be part of the Guidewire DevTrails 2026 Hackathon, a platform that challenges us to solve real-world problems using technology. The problem statement itself is both meaningful and complex—reimagining insurance for India’s growing gig economy.

From the very beginning, this journey has pushed us to think beyond conventional solutions and explore how AI can be used to create tangible impact.

2. Our Project — QuickShield

The Problem We're Solving:

India’s platform-based delivery partners (Zomato, Swiggy, Zepto, Amazon, Dunzo etc.) are the backbone of our fast-paced digital economy. However, external disruptions such as extreme weather, pollution, and natural disasters can reduce their working hours and cause them to lose 20–30% of their monthly earnings. Currently, gig workers have no income protection against these uncontrollable events. When disruptions occur, they bear the full financial loss with no safety net.

Our solution, QuickShield, is an AI-powered parametric insurance platform designed for delivery partners working with

->Q-Commerce PLATFORMS LIKE Zepto, Blinkit

These workers face frequent income disruptions due to external factors such as extreme weather, traffic congestion, or local strikes. Currently, there is no structured way to protect their earnings.

QuickShield addresses this by introducing a parametric micro-insurance model that operates on a weekly basis. Instead of manual claims, the system automatically detects disruptions using real-world data such as weather conditions and delivery activity.

If a disruption is confirmed, the platform ensures that workers receive a guaranteed portion of their expected daily income. This creates a safety net that is both efficient and fair.

While working on this project, we realized how critical it is to balance automation with fairness. Designing systems that can differentiate between genuine disruptions and normal fluctuations has been one of the most insightful parts of our experience.

3. The Pressure of the Hackathon Environment

One of the most intense aspects of this hackathon is the constant anticipation of updates. Every day, we find ourselves checking the dashboard for new announcements.

A particularly challenging element is the concept of a “market crash event.” This is an unexpected scenario introduced during the competition, where teams are given only 24 hours to adapt their solution. Failure to comply results in penalties.

This uncertainty adds a layer of pressure, pushing us to design our system to be flexible, scalable, and ready for sudden changes.

4. Working as a Team

Our team has been navigating this journey together, constantly discussing ideas, refining our approach, and supporting each other through challenges.

There have been moments of confusion and iteration, especially while aligning technical design with real-world constraints. However, these discussions have helped us build a stronger, more practical solution.

The collaborative process has been just as valuable as the technical learning itself.

5. Learning Along the Way

A few days ago, we participated in a quiz focused on Artificial Intelligence, Machine Learning, and core insurance concepts.

This experience reinforced our understanding of the domain and highlighted how important it is to combine technical knowledge with industry-specific insights. It also helped us validate many of the ideas we are implementing in QuickShield.

6. Looking Ahead

As we move forward, we remain committed to building a robust and thoughtful solution. This hackathon is not just about completing a project—it is about learning, adapting, and creating something meaningful.

We are determined to stay focused and see this through.

Our team, Chai_Pe_Charcha, is confident that our efforts will be fruitful.

DEV Community: Panthadeep Majumder