Skip to content

Harshit Kumar

  • Home
  • Projects
  • Blog
  • Poems
← Blog

Natural Language Processing

All Categories →

NLP from the ground up: word embeddings, attention mechanisms, transformers, text classification, and the building blocks of modern language understanding.

6 posts

  • Retrieval Augmented Generation (RAG) Chatbot for 10Q Financial Reports
    LLM

    Retrieval Augmented Generation (RAG) Chatbot for 10Q Financial Reports

    Building a RAG-based chatbot for 10Q financial reports to reduce LLM hallucinations by grounding answers in retrieved document context.

    Apr 26, 2024 · 5 min read
  • Attention
    Deep Learning

    Attention

    The attention mechanism in sequence-to-sequence models, how it allows the decoder to focus on relevant parts of the input at each step.

    Mar 08, 2019 · 3 min read
  • Backpropagation Through Time
    Deep Learning

    Backpropagation Through Time

    A mathematical deep dive into how gradients are computed in RNNs via Backpropagation Through Time (BPTT), explaining vanishing gradient origins.

    Feb 22, 2019 · 3 min read
  • Image captioning using encoder-decoder
    Deep Learning

    Image captioning using encoder-decoder

    Building an image captioning system using a CNN encoder and RNN decoder based on the Show and Tell architecture.

    Jan 11, 2019 · 2 min read
  • The gradient problem in RNN
    Deep Learning

    The gradient problem in RNN

    Why vanilla RNNs suffer from vanishing and exploding gradients, and how this limits their ability to capture long-range dependencies.

    Jan 04, 2019 · 3 min read
  • word2vec: The foundation of NLP
    Deep Learning

    word2vec: The foundation of NLP

    How word2vec represents words as dense vectors by learning from context, solving the limitations of one-hot encoding for NLP tasks.

    Jul 27, 2018 · 4 min read
Harshit Kumar 2026 About this site Creative Commons License