π Hi, I'm Lokesh!
I'm a curious and self-driven learner from India with a deep interest in Artificial Intelligence, Open Source, and technology for social impact. I use ChatGPT and other AI tools dailyβnot just to get answers, but to truly understand how they work under the hood. This curiosity has led me to explore model training, alignment, dataset creation, and the broader ethical implications of AI.
π My Interests
-
π€ Building AI models from scratch (not just using pre-built tools) -
π Making AI accessible in Indian languages & dialects - π§ Learning how models think, train, and follow alignment rules
π Highlight
-
π₯ Winner of CodeDay Hyderabad hackathon β led a team of strangers, learned full-stack basics in 24 hours, and built a winning website from scratch in 11th grade.
π§© What I'm Currently Doing
- Studying how Large Language Models (LLMs) are trained and aligned
- Exploring Swecha's projects like Swecha Voice, which make AI accessible in Telugu
- Learning Python, Deep Learning, and NLP
- Preparing to contribute to open-source Indic language AI projects
π§ Projects
Built a full-fledged weather application that fetches and displays real-time weather data using APIs. It includes a clean UI and responsive design using HTML, CSS, JavaScript, and integrates live weather updates via OpenWeatherMap API.
gthub link-https://github.com/Lokeshrao69/Weather-APP-1
π Telugu Corpus Website
Tech Stack: Astro Β· React Β· TypeScript Β· Java Β· CSS
Description: The Telugu Corpus App is a full-stack web application designed to support and promote the preservation and enrichment of the Telugu language. It empowers users to actively contribute to a growing Telugu digital corpus by enabling them to record voice clips, upload images, share videos, and submit written text related to their favorite topics, ranging from culture and literature to personal experiences and regional knowledge.
Key Features:
Purpose: This project aims to serve as a foundational tool for building open-source Telugu language resources, including speech-to-text engines, translation models, and academic linguistic studies.
Status:
Project: LangBuddy β AI-Powered Language & Coding Learning App*
(Team Thunder Buddies | Streamlit, OpenAI GPT-3.5, Python)
Role: Ai & Full-Stack Developer
Key Contributions:
- Designed full-stack architecture using Streamlit, integrated OpenAI API for dynamic explanations, and developed image-capture modules with PIL/OpenCV.
- Engineered prompt templates for GPT-3.5 to generate coding problems, writing feedback, and conversational simulations.
- Implemented user state management (st.session_state) for progress tracking and built interactive games (Hangman, Word Scramble).
- Conducted user testing (10+ participants), iterated based on feedback, and planned a 6-week roadmap for voice input and RAG enhancements.
Outcome: Launched on Hugging Face Spaces with 9/10 user satisfaction; praised for bridging language and CS education.
Skills Demonstrated:
Python Β· Streamlit Β· OpenAI API Β· Prompt Engineering Β· PIL/OpenCV Β· UX Design Β· Agile Development
π οΈ Skills
-
π» Programming: C++, Python, C -
π Web: Full Stack Development (HTML, CSS, JavaScript, Node.js, Express, MongoDB)