Hey there I'm

Harsha Vardhan Reddy Emani

I build

AI Research & Systems Engineer specializing in HPC and LLM optimization. Speaker @ SCI 2025. Ex-Intern @ C-DAC. Building production systems used by 5,000+ users.

View Projects Get in Touch

About

Who I Am

I'm Harsha Vardhan Reddy Emani, an AI Research & Systems Engineering student specializing in High-Performance Computing and Large Language Model optimization. Based in Guntur, Andhra Pradesh, India.

Hands-on experience fine-tuning and building models from scratch across AMD MI300X, Intel Gaudi3, and NVIDIA L40S GPU platforms. Speaker at SCI 2025 delivering tutorials on LLM mathematical foundations.

Core Member & Platform Manager at AIHUB VVIT — deployed 5+ production projects, organized hackathons, and mentoring peers in applied AI.

5K+ Users Served

5+ Deployments

8.7 GPA

What I Do

LLM Development & Optimization
Building LLMs from scratch, fine-tuning with LoRA, multi-GPU training
High-Performance Computing
CUDA, ROCm, OpenMP, GPU-accelerated AI pipelines
RAG & Multi-Agent Systems
LangGraph, LangChain, Vector DBs, FastAPI deployment
DevOps & Production Deployment
Docker, Linux servers, scalable backend architecture

Highlights

Featured Projects

Production-ready systems and ML solutions

LLM

Medical LLM & SLM Development

Built GPT-2–style Small Language Model from scratch with cross-hardware multi-GPU training on AMD MI300X systems using DDP and gradient accumulation.

Custom Causal Self-Attention, LayerNorm, Transformer blocks
8-GPU AMD MI300X distributed training with DDP
Fine-tuned 7B medical LLM using LoRA for parameter-efficient adaptation
Mixed-precision strategies for memory efficiency
Verifiable medical reasoning tasks with SFTTrainer

PyTorch Hugging Face ROCm/CUDA DeepSpeed LoRA

View Project →

RAG

SocialGPT — Production Multi-Agent College Intelligence Platform

Production multi-agent RAG platform serving 5,000+ students, parents, faculty, and administrators with stateful, role-aware academic intelligence.

LangGraph orchestrator with Claude as planner/synthesizer coordinating 8 specialized sub-agents
End-to-end RAG over 18 years of institutional PDFs (~18,000 chunks) using OCR, embeddings, and HNSW vector search
RBAC enforced at API middleware and tool layers with multilingual responses (English, Telugu, Hindi, Tamil)
Integrated MS Graph API, Edunxt ERP automation, and real-time web search via tool-driven workflows
Production observability with Prometheus, Grafana, Loki, and LangSmith plus anti-hallucination safeguards

Python LangGraph FastAPI MongoDB SurrealDB React Docker

View Project →

Production

TrackCoders – Student Analytics Platform

Production-grade campus-wide analytics platform used by 5,000+ students and faculty with real-time dashboards and progress visualization.

Deployed for 5,000+ students and faculty campus-wide
FastAPI + MongoDB backend with React frontend
Linux server deployment with Docker
Real-time dashboards and progress visualization
Optimized for high traffic and low-latency access

FastAPI MongoDB React Docker Linux

View Project →

Featured

SmartJob V2 - Fresher Jobs Platform

Comprehensive job and internship platform designed for freshers and students with real-time job updates from LinkedIn, Indeed, Naukri, and Internshala.

Advanced job search with location, experience, and remote filters
Real-time web scraping from multiple job portals
Integrated resume builder with multiple templates
Career guidance and interview preparation resources
Fast loading with intelligent caching and background updates

React TypeScript Tailwind CSS Vite Framer Motion

View Project →

Prompt Classification ML

Advanced machine learning system for hierarchical prompt classification with a web application for real-time inference.

Hierarchical classification model with superior accuracy
TF-IDF vectorization for robust text feature extraction
Data augmentation with expanded prompt dataset
Interactive Streamlit web application
Model persistence with pickle for efficient inference

Python Scikit-learn Streamlit TF-IDF

View Project →

Mobile

FarmConnect - Agricultural Platform

Modern Flutter-based cross-platform mobile application for connecting farmers with markets, featuring responsive design and comprehensive theming.

Cross-platform mobile application (Android & iOS)
Comprehensive theming system (light/dark modes)
Responsive design using Sizer package
Advanced routing system for seamless navigation
Production-ready modular architecture

Flutter Dart Material Design

View Project →

Deep Learning

Crop Price Predictor

Deep learning system for forecasting agricultural crop prices using Keras neural networks with automated data preprocessing pipeline.

Sequential Keras model with multiple dense layers
Automated preprocessing with feature engineering
Label encoding and standard scaling for data
Model persistence with preprocessing components
R2 score and MAPE evaluation metrics

TensorFlow Keras Python Pandas

View Project →

Time Series

Network Traffic Prediction

Time-series forecasting for network traffic patterns using dual approach: traditional ARIMA and modern LSTM neural networks.

Dual approach: ARIMA and LSTM models
Comprehensive EDA for traffic patterns
⚙️ Automated ARIMA parameter selection with pmdarima
Sequential LSTM for advanced prediction
RMSE and MAE performance evaluation

TensorFlow LSTM ARIMA Python

View Project →

Challenge

Amazon ML Challenge - Smart Pricing

ML solution for optimal product pricing in e-commerce, predicting prices for 75K products using multi-modal approach with text and images.

Price prediction for 75K+ products
Multi-modal: Text (catalog) + Image features
SMAPE evaluation metric optimization
Feature engineering for text and image data
Complete methodology documentation

Python ML Challenge Multi-modal

View Project →

Automation

Automated Job Market Analysis

Comprehensive web scraping solution for automated job market analysis using multiple Python libraries for data extraction.

Multi-library: BeautifulSoup, Scrapy, Selenium
Structured data extraction from job portals
Robust error handling and data validation
Efficient data storage and processing pipelines

Python BeautifulSoup Selenium Scrapy

View Project →

More Work

All Projects

Live data from GitHub

Fetching projects from GitHub…

Journey

Experience

December 2025
Technical Speaker — SCI 2025 Conference

International Conference

Selected to deliver a 3-hour tutorial titled "Mathematical Foundations for Large and Small Language Models" to AI researchers and engineers. Conducted live deployment and benchmarking across Intel Gaudi3, AMD MI300X (ROCm), and NVIDIA GPUs, comparing throughput, latency, and memory efficiency. Designed curriculum spanning Transformer internals, attention, positional encodings, and scaling laws.

Public SpeakingLLMsGPU Benchmarking
April 2025 – June 2025
HPC Intern

Centre for Development of Advanced Computing (C-DAC), Pune

Contributed to the HPC sub-project "High Performance Data Management Using Berkeley DB for Telecom Traffic Data". Designed and optimized graph-based C/C++ data structures for large telecom networks with millions of nodes and edges. Leveraged Berkeley DB and OpenMP for parallelized large-scale data access, with hands-on exposure to HPC clusters, profiling tools, and scalable visualization workflows.

HPCBerkeley DBOpenMPC/C++
2023 – Ongoing
Core Member & Platform Manager — AIHUB VVIT

AI/ML Community Organization

Lead AI/ML project development, mentor 100+ peers in applied machine learning, and organize 3 to 4 student hackathons per year. Architected, deployed, and maintain the official AIHUB VVIT website and backend infrastructure to ensure high availability for the campus AI community.

LeadershipDevOpsMentorshipCommunity
Aug 2023 – Present
B.Tech — Artificial Intelligence & Machine Learning

Vasireddy Venkatadri University (VVITU), Guntur

GPA: 8.7/10. Specializing in HPC and LLM optimization. Hands-on experience fine-tuning and building models from scratch across AMD MI300X, Intel Gaudi3, and NVIDIA L40S GPU platforms. Member of ACM student chapter.

AI/MLHPCLLM OptimizationACM Member

Credentials

Certifications & Achievements

Achievements

Technical Speaker — SCI 2025 Conference

3-hour tutorial: "Mathematical Foundations for Large and Small Language Models"

View Proof

Finalist - 5G Innovation Hackathon 2025

Recognized for innovative solution in 5G technology

View Proof

Ex-Intern @ C-DAC Pune (HPC Group)

Berkeley DB · OpenMP · Performance Profiling

View Proof

Core Member & Platform Manager — AIHUB VVIT

Deployed 5+ production projects · Organized 3-4 hackathons

ACM Student Chapter Member

Seminars, coding events, and tech talks participation

Oracle Cloud Certifications

•

OCI 2025 Certified Generative AI Professional

Oracle · Issued Oct 2025 · Expires Oct 2027

OCI GenAI RAG LLMs

View Proof

•

OCI 2025 Certified AI Foundations Associate

Oracle · Issued Aug 2025 · Expires Aug 2027

AI OCI ML Foundations

View Proof

Development & AI Certifications

•

Complete Flutter Development Bootcamp with Dart

Udemy · Issued Oct 2025

Flutter Dart Mobile Dev

View Proof

•

Programming in Java

NPTEL Online Certification · 2024

Java OOP Multithreading

View Proof

•

Generative AI by Google Cloud

Google Cloud Skill Boost (L4G) · 2024

LLMs Prompt Engineering Foundation Models

View Proof

Stack

Skills & Technologies

Languages

Python
C / C++
Java
JavaScript
Dart
SQL
HTML / CSS

AI / ML Frameworks

PyTorch
Hugging Face
LangChain / LangGraph
Scikit-learn
DeepSpeed
LoRA / DDP

Web & Mobile

React
FastAPI
Node.js / Express
Flutter
Scrapy
Tailwind CSS

Databases

MongoDB
MySQL
Oracle SQL
Berkeley DB
Vector Databases

GPU & HPC

CUDA
AMD ROCm
Intel Gaudi3
OpenMP
AMD MI300X
NVIDIA L40S

DevOps & Tools

Docker
Git / GitHub
Linux (Ubuntu)
Server Deployment
VS Code
Jupyter

Say Hi

Get In Touch

Open to opportunities, collaborations, and interesting conversations.

Email

GitHub

@harsha4261

Harsha Vardhan Reddy Emani

LeetCode

Eharsha4261

Harsha Vardhan Reddy Emani

Who I Am

What I Do

Featured Projects

Medical LLM & SLM Development

SocialGPT — Production Multi-Agent College Intelligence Platform

TrackCoders – Student Analytics Platform

SmartJob V2 - Fresher Jobs Platform

Prompt Classification ML

FarmConnect - Agricultural Platform

Crop Price Predictor

Network Traffic Prediction

Amazon ML Challenge - Smart Pricing

Automated Job Market Analysis

All Projects

Experience

Technical Speaker — SCI 2025 Conference

HPC Intern

Core Member & Platform Manager — AIHUB VVIT

B.Tech — Artificial Intelligence & Machine Learning

Certifications & Achievements

Achievements

Technical Speaker — SCI 2025 Conference

Finalist - 5G Innovation Hackathon 2025

Ex-Intern @ C-DAC Pune (HPC Group)

Core Member & Platform Manager — AIHUB VVIT

ACM Student Chapter Member

Oracle Cloud Certifications

OCI 2025 Certified Generative AI Professional

OCI 2025 Certified AI Foundations Associate

Development & AI Certifications

Complete Flutter Development Bootcamp with Dart

Programming in Java

Generative AI by Google Cloud

Skills & Technologies

Languages

AI / ML Frameworks

Web & Mobile

Databases

GPU & HPC

DevOps & Tools

Get In Touch

Email

GitHub

LinkedIn

LeetCode