π Led a 3-member team to 2nd Place (out of 800) at Google DSC IIT Kanpurβs βThe Return Journeyβ Hackathon (2024).
π‘ Passionate** about problem-solving, scalable cloud architectures, and serverless microservices.
π₯ Focused** on high-performance APIs, cloud automation, and cost-effective infrastructure solutions.
π SwiftVault - Enterprise File Vault System π Live at: swiftvault
π Timeline: 8 Days π Built for BalkanID 80 LPA Full-Time Hiring Challenge (September 2025) (In 6 Days as Solo Developer -> Ideation + Architecture + Coding + Testing + Cloud Deployment + Extras -> 2 Days Edge Native - Serverless Optimization)
Built a production-grade secure file vault with intelligent deduplication, custom JWt implementation with token generation, claim-based access control, token verification, refresh token, advanced sharing, and comprehensive analytics in just 6 days - then revolutionized its architecture in 2 more days to achieve near-zero costs.
π Architecture Evolution
Pain Points: Idle costs, single point of failure, regional deployment
Benefits: Zero idle costs, global distribution, pay-per-use, independent scaling
- π° Cost Optimization: Reduced from $204/year β $0 (Year 1), $108/year (Year 2+) = 47% permanent savings
- β‘ Performance: <100ms metadata, <500ms uploads with edge caching
- π Security: RS256 JWT, bcrypt hashing, MIME validation, SHA-256 deduplication
- π Efficiency: 30-40% storage savings through intelligent deduplication
- π Global: Distributed edge architecture with <50ms response times
- π Scale: Designed to handle 10K+ req/sec with zero-cost scaling
Backend: Go, GraphQL, PostgreSQL, Redis, MinIO Frontend: Next.js 14, TypeScript, Tailwind CSS Cloud: AWS (t2.micro), Vercel, Cloudflare Workers/KV/R2, Neon PostgreSQL DevOps: Docker, Nginx, SSL/TLS, GitHub Actions
- π§ Intelligent Deduplication: SHA-256 hash-based duplicate detection
- π Zero-Client Architecture: All client side sent data are cross-verified before use.
- β‘ Real-time Processing: Async uploads with live progress tracking
- π― Advanced Search: Multi-criteria filtering with tag support
- π Admin Analytics: Comprehensive system monitoring
- π Flexible Sharing: Private user sharing + public link sharing
- π Rate Limiting: Redis-based throttling (2 req/s user, 10 req/s IP)
β Monoliths are expensive for unpredictable traffic β Serverless architecture enables true pay-as-you-go economics β Edge-native is the future: low latency + zero idle costs β Free tiers are powerful when architected correctly
π― VersionCV - AI-Powered Resume Optimization Platform (Beta Launched) π Live at: www.versioncv.com
π¨βπ» Role: Founder & Solo Developer ποΈ Registered: MSME β Govt of India π Timeline: June 2025 β Present
Transforming how professionals manage and optimize their resumes with AI driven resume rewriting, ATS optimization, JD Matching, version control, personal vaults, resume to portfolio.
-
π€ AI Resume Optimization Agent
- Gemini, Chatgpt, claude, mistral used in fallback manner
- Job-specific structured prompts
- Section-wise ATS optimization
- Can rewrite whole resume under 30 seconds.
- Preserved Links original data accuracy.
- Hallucination Control using algorithmic prompt.
- Token Optimization for cost and ratelimits.
-
π Instant ATS Analyzer
- Real-time scoring and feedback
- Keyword optimization suggestions
- Format compatibility checking
- ATS parsability detection
- Complex layout, missing section, word density etc. detection correction.
-
π Custom VersionCV Editor (zero dependency)
- React Powered WYSIWYG Editor with ATS Scanner Integrated.
-
ποΈ Smart Resume Management
- Role-based version control
- Shareable, CDN-accessible links
- Cloudflare R2 storage with versioning
- Clone and edit.
- Can do resume correction even after link submission replace original at same link.
- Track Analytics for each of the resume file, role, vault.
-
π Enterprise Security
- Firebase JWT authentication on each request. (1ms full claims verification)
- Email verification.
- Zero Client Trust model.
- Main DB and buckets accessible only by internal api( no dns ) accessible through rpc calls only.
- API interlocking using different secrets for each and are present only on the cloudflare secret manager.
- Multiple layer of DDOS mitigation strategy, Cloudflare, internal Rate Limiter with 4 fallbacks.
Architecture: Fully decoupled, edge-native serverless, multi-cloud microservices Deployment: Google Cloud Run, Vercel Edge, Cloudflare Workers, D1/R2/Queue/KV Cache/Durable Objects/Cloudflare CDN Performance: <100ms latency, 10K+ req/sec capability, 99.9% uptime Security: DDoS protection, bot verification, Firebase JWT auth, API interlocking using internal secrets, Ratelimiter with 4 layered fallback mechanism
- Primary DB: Cloudflare D1 (SQLite) with fallback protection
- Caching: KV Cache Cloudflare CDN with real-time invalidation (>99% hit rate)
- Storage: Cloudflare R2 for CDN-accessible, versioned resumes
- Optimization: Batch-based DB updates preventing race conditions
Running on βΉ900/year (domain only) β entire infrastructure on zero-cost tiers!
Project | Description | Tech Stack | Live Link |
---|---|---|---|
Atithidev Couch Surfing | Hospitality exchange platform with real-time matching | Node.js, MongoDB, Socket.io | Visit |
SonarGuard Home Security | Radar-based security system with real-time alerts | Python, Arduino, Flask | Visit |
Atithidev DB API | Scalable REST API for hospitality platform | AWS Lambda, DynamoDB | Visit |
Sentiment Analyzer | NLP-based sentiment analysis microservice | Python, AWS Lambda, NLTK | Visit |
I'm open to working on high-impact projects involving:
- ποΈ Complex distributed architectures
- βοΈ Cloud cost optimization and automation
- β‘ High-performance API design
- π Serverless and edge computing
- π Security-first system design
Let's build something amazing together! π
Building production-grade secure systems that scale efficiently and cost-effectively