I am a Full-Stack Developer and ML Engineer focused on shipping production systems. I build real-time web applications using Next.js, TypeScript, and AWS, and develop computer vision pipelines using YOLOv8, PyTorch, and OpenCV for video analytics. At my current internship with AI Secure Solutions, I contribute to an enterprise video analytics platform for real-time CCTV monitoring. Previously at Chitralai, I built an internal analytics dashboard using live AWS data and shipped SEO improvements that increased organic traffic by 20%. I've also built and deployed a real-time multiplayer event platform for my campus using Next.js and Supabase. Beyond individual contribution, I serve as Vice President of the GitHub Community at GITAM, where I lead technical events and workshops for over 200 students. I care about taking end-to-end ownership of features, optimizing performance, and delivering working software under tight deadlines.
SOFTWARE ENGINEER INTERN
AI SECURE SOLUTIONS
- Contributed to the development of an enterprise Video Analytics (VA) platform used for real-time CCTV monitoring, event detection, and security operations.
- Participated in testing, validation, and performance evaluation of AI-powered video analytics features including object detection, tracking, and event-based alerting.
- Collaborated with cross-functional teams on product architecture, UI/UX improvements, deployment workflows, and technical documentation.
FULL STACK INTERN
CHITRALAI
- Built an internal analytics dashboard using live AWS data sources (DynamoDB, S3) to track subscriptions, uploads, organizations, users, and platform activity.
- Developed interactive visual reporting modules including pie, bar, and trend charts for faster business monitoring and decision support.
- Built a role-based blog CMS with admin/editor access, rich text editing, image uploads, draft/publish workflows, and SEO controls for non-technical teams.
- Implemented SEO improvements across public pages, contributing to a 20% increase in organic traffic over two weeks.
The multi-agent pipeline routes through 4 specialized LangGraph nodes. The Risk Profiler uses a CatBoost classifier (100% accuracy) to segment users into risk tiers based on 25 driving history parameters. High-risk quotes trigger the Conversion Predictor, heavily utilizing SMOTE to counter a massive class imbalance (22% bind rate). This node runs real-time explainability logic, generating SHAP, LIME, Anchors, and DiCE values in <200ms per quote. Finally, the Premium Advisor leverages Groq Llama 3.3 70B to synthesize the ML outputs into a human-readable justification, routing the final decision for auto-approval or human underwriter escalation.
Built from scratch using a local Pygame engine replica, this AI utilizes NEAT (NeuroEvolution of Augmenting Topologies) to literally evolve a brain capable of flawlessly playing the Chrome Dino game. Unlike standard neural networks where only weights are updated, NEAT evolves both the weights and the physical structure of the network via genetic algorithms. The population started with simple 12-input, 2-output nodes and dynamically grew hidden nodes to handle complex decisions like ducking under varying bird heights. Implemented a rigid fitness function rewarding survival while heavily penalizing incorrect actions (e.g., -30 points per frame for failing to duck). The final optimized network easily surpasses a 10,000 score limit.
Bypasses standard, slow DOM color picking algorithms by rendering images into a hidden HTML5 Canvas and directly mapping the raw Uint8ClampedArray pixel buffer. Once extracted, RGB vectors are mathematically translated into the perceptually uniform LAB color space. This prevents dark or muddy colors from mathematically dominating the palette generation compared to standard Euclidean RGB distance. The CIE94 delta-E formula is then applied to calculate human-visual distance. A strict filtering pass strips out <0.5% pixel noise anomalies before a K-Means clustering algorithm cleanly groups the top 5 dominant swatches in under 200ms.
B.Tech
GITAM deemed to be university, Hyderabad
AWS Academy Graduate - Cloud Foundations - Training Badge
Amazon Web Services Training and Certification
Dec 2025MongoDB
Jun 2025
Attendee · Oct 24, 2026