Building high-performance web systems.
I ship AI-powered products & full-stack systems that solve real health and business problems. Currently scaling CardioNerve and Thinknode AI.
// 01 — Selected Work
Things I've built — from idea to deployment. Focused on impact and scale.
Problem: Many hackathon teams fail due to unclear planning, poor time management, and confusion during the critical first few hours.
Solution: Developed an AI copilot that instantly generates structured roadmaps, workflows, tool stacks, and presentation guides.
HackMate is an AI-powered hackathon copilot built to streamline project planning, workflows, AI prompts, and demo prep, turning raw ideas into hackathon-ready workflows within minutes.
Problem: Delayed diagnosis in cardiovascular diseases leads to high mortality rates in remote areas.
Solution: Built an AI-powered risk assessment platform that analyzes patient vitals in real-time using predictive modeling.
As Project Lead, I orchestrated the development of this clinical-grade system. Features real-time data ingestion, validation pipelines, and an interactive analytics dashboard for actionable clinical insights.
Problem: Farmers lack real-time soil data, leading to fertilizer waste and sub-optimal yields.
Solution: Developed an AI platform with IoT sensor integration for predictive soil health monitoring.
AI-powered agricultural intelligence platform designed to optimize crop yield and provide predictive insights for smart farming. Helped 50+ farmers improve yields by an average of 10%.
Problem: Counterfeit drugs cause thousands of deaths annually due to lack of verification.
Solution: Built an AI verification platform for chemical composition validation and batch tracking.
Pharmaceutical verification platform designed to detect counterfeit and substandard medicines. Enables batch tracking and compliance monitoring through intelligent data analysis.
Problem: LLM prompts in UX research are inconsistent, unvalidated, and unreproducible — making research findings unreliable.
Solution: Built a reproducible LLM prompt framework with audit hashes, strict schema validation, and a CLI — enabling traceable, replicable UX research workflows.
A structured prompt engineering framework built for Google Summer of Code 2026 @ RUXAILAB. Enforces reproducibility in AI-assisted UX research through deterministic prompt versioning, audit trail hashing, and CLI-based validation pipelines.
// 01.5 — Community
Active engineering contributions to open-source and ecosystem codebases this year.
Refactored the core application brand component in the main header to act as an active, clickable home link, streamlining user navigation across CP Contest Editorials (editorial.coduck.io).
Designed and implemented high-visibility colored headers on search result cards, establishing strong visual hierarchy and categorizations for CP editorials.
Investigating a performance bug where apply_settings() queries the database on every invocation instead of caching — causing unnecessary load on the ESP platform at scale.
Debugging a crash in the List Generator where the old selector panics on stale or invalid filter IDs, causing silent failure during student list generation workflows.
Resolving a silent exception suppression bug in esp/web/views/csrf.py where exceptions are caught without logging — masking critical CSRF-related errors in production.
// 02 — Experience
Where I've been and what I've built along the way.
Selected as a Google Gemini Ambassador — representing and promoting Google's AI ecosystem by engaging with developer communities, exploring generative AI technologies, and contributing to AI-driven innovation through workshops, hackathons, and community initiatives.
Leading the development of an AI-driven cardiovascular risk assessment system. Key achievements: Led team of 4 engineers, reduced prediction latency by 30%, and achieved 99% uptime during pilot testing.
Developing scalable fintech systems. Metrics: Optimized SQL queries reducing load time by 40%, built 5+ responsive dashboard modules, and handled 10k+ financial data points daily.
Worked on AI consulting projects. Impact: Developed 3 ML models for automated data labeling, improved classification accuracy by 15%, and processed 1M+ data rows.
Deep dive into AI/ML fundamentals, data structures, algorithms, and distributed systems — combined with real-world project-based learning and industry-focused training.
// 03 — Impact
Data-driven proof of development scale and system reliability.
// 04 — Beyond the Classroom
Hackathons, achievements, certifications — the stuff textbooks don't cover.
Selected among the Top 70 out of 75,000+ participants nationwide. Developed an innovative AI solution targeting regional language accessibility using GPT-4o fine-tuning. Advanced to the national finals held at Bengaluru.
Ranked in the Top 69 out of 15,000+ teams globally — placing in the top 0.5% of all participants. Competed against international teams with a high-impact AI-powered solution, advancing to the finalist round.
Founded and launched an AI automation agency targeting SMBs. Built local-first LLM agents for customer support and automated 10+ workflow pipelines for initial clients. Currently scaling operations.
Secured **2nd Prize** among 2000+ members (500+ teams) in an intense campus offline hackathon. Developed a decentralized identity system for student records within 24 hours of starting the program.
Deep knowledge of cloud architecture, serverless (Lambda), and S3 optimization.
Mastery in building intelligent chatbots and implementing machine learning models using IBM Watson.
Foundational knowledge of Azure AI services including Computer Vision and NLP.
Training in generative AI architectures and Stable Diffusion model implementation.
Certified Specialist in data analysis, visualization, and PowerBI intelligence.
Developer adding next achievement... stay tuned
// 05 — About
// 06 — Contact
Take a break from the code. Neutralize the corrupted data nodes. Don't hit the mainframe boundaries.