← Back to portfolio

Rakshith Reddy Aredla

Full-Stack + AI Developer

📍 Hyderabad, India 💬 Full-stack · AI · RAG · Agent memory
Download PDF ↗

Summary

Full-stack + AI developer in Hyderabad. Mostly TypeScript and Python — Next.js and Supabase on the front, FastAPI when there's a model involved. I build small tools that are small enough to finish and real enough to use, shipping everything end-to-end. Currently exploring retrieval and the parts of agent memory that turn out to be harder than they sound.

Technical Skills

TypeScriptJavaScriptPython Next.jsNode.jsFastAPI FlaskPyTorchReact Native ML KitSupabasePostgres Razorpay APIStreamlitTailwind CSS GitGitHub ActionsVercel

Selected Projects

Full-stack · Live

Arkaira — Online Flower Shop

Realtime stock that decrements the moment someone pays, Razorpay checkout, admin dashboard. Next.js + Supabase + TypeScript.

AI · On-device

Netra — Driver Drowsiness & Distraction Detection

Real-time AI safety monitoring running on-device with React Native + ML Kit — no server required.

AI · Agent memory

RecallDesk — Support Copilot That Remembers

Customer support agent with persistent memory — recalls prior tickets and what actually fixed them.

AI · Health

Anchor — Caregiver Memory Navigator

Memory-powered navigator for family caregivers — meds, labs, doctor's instructions in one place. FastAPI + Streamlit + LLM.

AI · RAG

CascadeGuard — RAG Pipeline Monitoring

Implementation of a published paper — observability for retrieval pipelines, tracing every stage to catch silent failures.

Developer tools

GitGlance · EnvGuard · SpamScope

GitHub profile scorer CLI, fail-fast typed env validation for Node, and a Naive Bayes spam classifier written by hand — no sklearn, no numpy.

Experience & Involvement

2026

HackWithHyderabad 3.0 — Participant

Built and shipped under hackathon pressure — designed, implemented and presented a working product within the event.

2026

AI/ML Coursework

Completed an AI/ML course covering deep learning, retrieval-augmented generation and applied ML — applied directly in shipped projects.

Learning Path

The beginning

First website · First Python

My first HTML/CSS page and the classic number guessing game. Where it all started.

Foundations

ML Exercises

CNN on MNIST (~99%), RAG chatbot, house-price regression (R² ≈ 0.80), movie recommender, iris classifier.

Built with real code, honest commits — © 2026