SolarPulse by Qash.stp2
Under the PM Surya Ghar Muft Bijli Yojana, India is witnessing a massive push toward residential rooftop solar. Yet, millions of homeowners stall at the very first step of their journey: discovery. Currently, if an Indian homeowner wants to know if their roof is viable for solar, and search for it, many new AI wrappers attempt to calculate solar capacities using Large Language Models (LLMs), leading to mathematical hallucinations regarding panel efficiency, carbon offsets, and payback periods.
SolarPulse: Accelerating India’s Rooftop Solar Transition with Privacy-First AWS Serverless AI
Team: Qash.stp2
Track: Track 03 - Waste and Energy (Bharat Builds Tour)
Links:
The Problem: Friction and Privacy in the Solar Funnel
Under the PM SuryaGhar Muft Bijli Yojana, India is witnessing a massive push toward residential rooftop solar. Yet, millions of homeowners stall at the very first step of their journey: discovery.
Currently, if an Indian homeowner wants to know if their roof is viable for solar, they are forced to use calculators that act as aggressive lead-generation tools. These platforms demand invasive Personally Identifiable Information (PII)—like exact phone numbers and street addresses—just to provide a generic estimate.
Worse, many new AI wrappers attempt to calculate solar capacities using Large Language Models (LLMs), leading to mathematical hallucinations regarding panel efficiency, carbon offsets, and payback periods.
The Solution: SolarPulse
SolarPulse is a privacy-first, serverless rooftop feasibility platform. By simply entering a monthly power bill, city, and roof type, users instantly receive:
- Deterministic System Sizing: Exact kW capacity, roof footprint, and generation metrics.
- Financial Engineering: Capital costs offset by exact PM Surya Ghar central subsidies, calculating net investment and payback years.
- AI Structural Advisory: An engineering-style installation advisory dynamically generated by Amazon Bedrock based on the user's specific roof type and location.
We deliver all of this in under 3 seconds, with zero personal data collection and zero sales friction.
AWS Architecture: 100% Serverless & Cloud-Native
To achieve global scalability and sub-second deterministic calculations, we built SolarPulse entirely on AWS, utilizing AWS Serverless Application Model (SAM) as our open-source Infrastructure-as-Code (IaC) foundation.
- AWS Amplify: Our React 18 + Tailwind CSS frontend is deployed via Amplify Hosting. This provides global edge caching, instant CI/CD from our GitHub repository, and custom branch deployments.
- Amazon API Gateway: Acts as our secure HTTP API front door, routing requests to our compute layer while providing automatic CORS management and payload size throttling.
- AWS Lambda: The brain of SolarPulse runs on Node.js 24 using the ARM64 Graviton architecture for maximized cost-to-performance efficiency. Lambda executes our deterministic solar physics equations and orchestrates upstream AI calls.
- Amazon DynamoDB: We utilize DynamoDB as a high-speed atomic counter to enforce strict API rate limiting. To protect user privacy, client IPs are salted and SHA-256 hashed before hitting the database. We utilize native DynamoDB Time-To-Live (TTL) to automatically permanently delete these anonymous hashes within 60 minutes.
- Amazon Bedrock: We leverage the Anthropic Claude Haiku 5.5 foundation model via Bedrock's unified API. Bedrock provides serverless generative intelligence without the overhead of provisioning GPU infrastructure.
Architecture:
Engineering Depth: The Strict Separation of Math and AI
One of our primary engineering philosophies was acknowledging the limitations of LLMs. LLMs are exceptional at language synthesis but inherently unreliable at deterministic math.
To solve this, SolarPulse implements a bifurcated processing pipeline:
- Audited by Lambda: The exact solar panel sizing, rooftop square footage requirements, carbon dioxide offsets, and PM Surya Ghar subsidy tiers (up to ₹78,000) are hard-coded into our Node.js rule engine (
solarFormulas.js). The math is absolute and executed locally in the Lambda execution environment.
- Synthesized by Bedrock: Once Lambda generates the authoritative metrics, it passes them as a strict payload to Amazon Bedrock. Bedrock's only job is to analyze the structural environment (e.g., a "Concrete Flat" roof vs. "Sloped Metal") and generate qualitative guidance, such as recommended tilt angles, mounting structures, and net-metering application steps.
High Availability & Graceful Degradation:
We engineered our Lambda handler with a resilient in-memory fallback mechanism. If Amazon Bedrock experiences an upstream timeout or a latency spike, the Lambda function gracefully aborts the AI invocation and instantly falls back to a deterministic, rule-based engineering advisory. This guarantees the user always receives their audit in under 3 seconds with zero downtime.
Security & Privacy First
In Track 03 (Waste and Energy), building trust is as important as building the technology.
- Zero PII: We do not collect names, emails, phone numbers, or GPS coordinates.
- Ephemeral Processing: All user inputs are processed strictly in-memory during the Lambda execution lifecycle.
- Anonymous Throttling: By combining SHA-256 cryptographic salts with DynamoDB TTL, we successfully built a robust DDoS and rate-limiting system that is structurally incapable of tracking users long-term.
Conclusion: Empowering Bharat
SolarPulse removes the barriers to clean energy discovery. By combining the deterministic computing power of AWS Lambda with the contextual intelligence of Amazon Bedrock, we have created an open-source tool that empowers Indian homeowners to transition to rooftop solar with confidence, transparency, and total privacy.
Explore the Project:
- GitHub (Open Source IaC & Code): AshQ-ghbi/SolarPulse-Qashfin
- Live Web App: Try SolarPulse Here
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article