Skip to content
View pavithra20august's full-sized avatar
2 followers · 0 following

Block or report pavithra20august

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
pavithra20august/README.md

Pavithra Vijayan

Machine Learning Engineer  |  10 Years in Production ML

LinkedIn YouTube Email

I build and deploy production ML systems — from fine-tuning LLMs to real-time forecasting pipelines. Currently at Arista Networks, working on generative AI, NLP, and time-series forecasting on GCP.

What I work with:

LangChain LangGraph RAG MCP Hugging Face LoRA/QLoRA PyTorch TensorFlow XGBoost scikit-learn GCP Vertex AI Docker FastAPI React

Featured Work

Project What it does Stack
LangGraph Stateful Agent Deep research agent with state machines, real-time SSE streaming LangGraph, FastAPI, React
CrewAI Crash Course Multi-agent system with 4 AI agents — Startup Idea Validator CrewAI, FastAPI, React
Multi-Agent LangChain Multi-agent orchestration for complex task execution LangChain, Python
Email Classification DistilBERT-based email classifier — 92% accuracy Transfer Learning, NLP
Hybrid LLM Classifier Hybrid ML + LLM classification system LLM, TypeScript
MCP Server Model Context Protocol server for AI tool integration MCP, Python
Trade Prediction Agent AI agent for financial market prediction ML, Python
NanoGPT GPT implementation from scratch — transformers demystified Deep Learning, Python
Talking to PDF RAG-based document Q&A with vector embeddings RAG, LLM, Python
Vertex AI Workflows ML deployment pipelines on Google Cloud GCP, MLOps

Experience Highlights

  • Arista Networks (2021–Present) — Building LLM-powered summarization, email classification (92% acc), time-series forecasting, and RAG-based recommendation systems
  • EXL Health (2019–2021) — NER systems for medical insurance risk adjustment, SDOH prediction models using WHO data
  • HCL Technologies (2017–2019) — Speech-to-Text on 11K hours of accent data (CNN/RNN), speaker diarization, Docker-deployed real-time inference
  • Published Author — "Become Data Scientist" on Amazon Kindle
  • Stanford Online — Statistical Learning: Applied Machine Learning

GitHub Activity

GitHub Stats

Top Languages

Popular repositories Loading

  1. Self-Evolve Self-Evolve Public

    3

  2. pavithraspodcast-files pavithraspodcast-files Public

    Source files and resources for Pavithra's Podcast — ML/AI educational content with 10K+ YouTube subscribers.

    Python 2 1

  3. Email-Classification Email-Classification Public

    Email classification pipeline using DistilBERT with transfer learning — 92% accuracy. Automated categorization with preprocessing pipelines.

    Jupyter Notebook 1

  4. K-Fold-cross-validation K-Fold-cross-validation Public

    K-Fold cross-validation implementation for robust model evaluation and hyperparameter tuning.

    Python 1

  5. Mullti-Agent-Langchain Mullti-Agent-Langchain Public

    Multi-agent orchestration system built with LangChain — coordinating multiple AI agents for complex task execution.

    Jupyter Notebook 1

  6. Hybrid-LLM-Classifier Hybrid-LLM-Classifier Public

    Hybrid classification system combining traditional ML with LLM-based approaches for improved text categorization accuracy.

    TypeScript 1