👋 Hi, I'm Ved! I’m🎖️ a Machine Learning engineer focused on harnessing technology to make significant impacts in the tech industry and beyond.
🚀 Young and ambitious developer on a mission to create impact through innovative AI solutions! Currently pursuing Minor in AI from IIT Ropar while diving deep into self-learning adventures.
🔮 Whether it's teaching machines to see, understand language, or solve complex problems, I'm your go-to person for turning AI dreams into reality! Let's build something amazing together!
- Languages: Python, C , javascript
- Frameworks & Libraries: Fast ai , Pytorch , Pandas , Matplotlib
- Tools & Platforms: GitHub, Docker, , Vercel, Neovim , Vscode , JupyterNoterbook , Aws
- Machine Learning Specialist: Proficient in statistical analysis, predictive modeling (Regression, Decision Trees, Random Forest), and advanced algorithms (CatBoost, SGD) with strong focus on optimization and accuracy.
- WED BOT: A bot who mimics my personality and talks like me (fun fact: It has 2 personalities). None is like the real me
- Pdf_summariser: Innovative RAG web app for PDF analysis using machine learning. Enables intelligent document summarization and information retrieval through advanced embedding and tokenization techniques. Transforms PDF content into actionable insights with AI-powered processing.
- LLM tokenizer: Bulding a tokenizer for llm from scratch :)
- Loan approval predictor: Utilized a huge dataset of loan defaults and performed data transformation and model training. Tried different methods to get 95%+ accuracy and deployed with AWS and implemented CI/CD pipelines
- Image classification Model: Trained a neural network from scratch in pure Python (no libraries) and achieved 80%+ accuracy on the MNIST dataset
- Soptify song recom: Optimized a recommendation system through A/B testing and advanced algorithms, boosting click-through rates by 60%, user satisfaction by 35%, engagement by 40%, and listening time by 25%.
- Fine tuning Guide: Fine-tuning the google/bert-base-cased language model using the SFTTrainer from TRL with memory-efficient settings like gradient checkpointing and 8-bit quantization.
Many in deployment many failed T_T"
-
v0 clone: tried making a v0 clone that generats websited was able to create a backed to get the code from gemini but showing it on frontend and running it in a container was not able to do that (upkilling it right now will make it someday pakka : )
-
Research-copilot: working on a chat bot that will help in research and provide relavent links citations and summarization of given paper made with RAG and llama models used vector DB storage nad a=Advanced Rag models
- FreecodeCamp: Wrorked as Teaching Assistance Utilized advanced teaching methods to facilitate hands-on coding projects, resulting in a 30% increase in student engagement and participation levels.
I’m looking forward to collaborating on projects that are at the intersection of technology and social good. Let’s connect! 🌍