HEMAKRISHNA.AI
Available for Hire
DATA SCIENCE & GENERATIVE AI ENGINEER

Engineered LLM Pipelines & Machine Learning Systems.

B.Tech Computer Science graduate specializing in production GenAI workflows, transformer fine-tuning (FLAN-T5), and scalable document intelligence architecture.

View Projects Recruiter Contact

FLAN-T5

Fine-Tuned Summarization Model

100%

Context Extension via Token Chunking

3+

Technical Articles Published on Medium

B.Tech CSE

Core Computer Science Engineering

Projects Portfolio

Real-world AI engineering applications, predictive models, and full-stack software.

FLAGSHIP GENAI PROJECT

DocMind AI — Document Intelligence Platform

A GenAI-powered platform designed to process long-context PDFs, generate abstractive summaries, and eliminate model context limits.

  • Model Fine-Tuning: Adapted FLAN-T5 on the GovReport dataset using PyTorch & Hugging Face.
  • Hierarchical Summarization: Token-based chunking with overlapping windows and batched iterative summarization[cite: 5].
  • Fact Preservation: Deduplication and entity verification rules to prevent hallucinations[cite: 5].
Python PyTorch Hugging Face FLAN-T5 Streamlit PyMuPDF
Pipeline Flow v1.0.4
01. Extraction PyMuPDF Parsing
02. Processing Overlapping Window Chunking
03. Inference Fine-Tuned FLAN-T5
04. Guardrails Deduplication Engine
05. Interface Streamlit Dashboard

Mental Tiredness Score Prediction

Machine learning application predicting fatigue scores based on lifestyle factors with database history tracking[cite: 5].

  • Trained linear & tree regression models evaluated via R², MAE, and RMSE metrics[cite: 5].
  • Integrated MySQL database for user prediction history logging[cite: 5].
  • Built an interactive Streamlit frontend for live parameter inputs[cite: 5].
Python Scikit-Learn MySQL Streamlit

Crypto Market EDA & Price Forecasting

Automated web scraping pipeline collecting real-time crypto market data for regression forecasting and exploratory analysis[cite: 5].

  • Scraped financial market data using BeautifulSoup & Requests[cite: 5].
  • Performed exploratory data analysis (EDA) with Seaborn correlation heatmaps[cite: 5].
  • Predicted short-term price movements using feature engineering[cite: 5].
Python BeautifulSoup Pandas Matplotlib

Bank Application & Transaction Engine

A full-stack banking application handling user authentication, transactions, and account management[cite: 5].

  • Designed relational database schemas with complex SQL joins and constraints[cite: 5].
  • Implemented CRUD operations for deposit, withdrawal, and transfer workflows[cite: 5].
  • Integrated input validation and transaction security checks[cite: 5].
Python SQL / MySQL OOP Streamlit

Technical Expertise

Core skills matched for entry-level Data Science & GenAI roles[cite: 5].

Generative AI & NLP

LLM Fine-Tuning FLAN-T5 Transformers RAG & LangChain Prompt Engineering Token Chunking

Machine Learning & DL

PyTorch TensorFlow Scikit-learn ANN / CNN / RNN LSTM & GRU

Core Languages & SQL

Python SQL / MySQL Pandas / NumPy Streamlit Git / GitHub

Technical Writing

Deep-dive technical breakdowns published on Medium[cite: 5].

From Words to Intelligence: The Complete Journey of NLP

Covers NLP evolution, embeddings (Word2Vec), Attention Mechanisms, Transformers, BERT, and GPT[cite: 5].

Read Article

Artificial Neural Networks & Machine Learning Fundamentals

In-depth breakdown of MLP architectures, activation functions, backpropagation, and hyperparameter tuning[cite: 5].

Read Article

Get In Touch

Actively open for Data Science, Machine Learning, and GenAI Roles[cite: 5].

Email Me

nidadavoluhemakrishna@gmail.com

Location

Hyderabad, Telangana, India[cite: 5]

LinkedIn

Connect Professionally

GitHub

View Code Repositories
Send Direct Email