EeshOS v1.0

ABOUT.EXE

🧑‍💻

EESH SAXENA

> CS Engineer & AI Researcher

📍 Gandhinagar, Gujarat, India

IDENTITY

Institution

IIIT Senapati, Manipur

Degree

B.Tech, CSE

Phone

+91 7976212108

Email

[email protected]

CONNECT

⌥ github.com/eeshsaxena 🔗 linkedin/eesh-saxena 📊 kaggle.com/eeshsaxena 🧩 codolio.com/Eeshsaxena_ ✉ [email protected]

BIO

CS undergrad at IIIT Senapati passionate about AI/ML, NLP, and competitive programming. Research intern at IIIT Vadodara (Reversible Data Hiding) and IIT Tirupati (Multi-Object Tracking). Building open-source NLP corpora and deploying edge AI on embedded hardware. Competitive programmer with 1500+ problems solved and global ranks across Codeforces, CodeChef & LeetCode.

eesh@os:~$ whoami

EDUCATION.EXE

DEGREE

Indian Institute of Information Technology Senapati

Bachelor of Technology in Computer Science & Engineering

📅 Oct 2023 - Present | 📍 Manipur, Imphal

🎓 B.Tech CSE (Active)

Focus Areas

AI / Machine Learning

Natural Language Processing

Computer Vision

Data Structures & Algorithms

Embedded Systems

Database Systems

Operating Systems

Reversible Data Hiding

RESEARCH PROFILE

Problems Solved

1,500+

across all platforms

Contests Participated

50+

competitive programming

Research Internships

2

IIIT Vadodara + IIT Tirupati

GitHub Repositories

51

public repos

eesh@os:~$ cat education.txt

EXPERIENCE.EXE

RESEARCH INTERNSHIPS

🏛 IIIT Vadodara, under Dr. Abhisek Paul

Research Intern (2026)

📅 2026 | 6-month internship

Secured internship to analyze Zhang (IEEE SPL 2011) framework on Reversible Data Hiding in Encrypted Images.
Studied and evaluated techniques from 30+ related research works to compare embedding strategies.
Developed the encryption, data-embedding, and extraction pipeline ensuring separability between data retrieval and image reconstruction.
Measured embedding capacity and fidelity using PSNR and embedding-rate metrics for publication.
Implemented embedding algorithms in Python with near real-time encryption and data hiding speeds.

🏛 IIT Tirupati (Online), SEVA Lab, under Dr. Chalavadi Vishnu

Winter Research Intern (2025)

📅 Winter 2025 | Online

Reproduced and analyzed the MOTIP (CVPR 2025) framework for Multiple Object Tracking.
Focused on transformer-based identity prediction without Hungarian-based association.
Examined spatio-temporal feature embeddings under occlusion and dense-scene conditions.
Applied temporal regularization to reduce identity switches in tracking scenarios.
Evaluated performance using MOTA, IDF1, and HOTA metrics after adjusting optimization parameters.

eesh@os:~$ ls ./experience/

PROJECTS.EXE

FEATURED PROJECTS

⬡ Conflict-Aware Graph RAG

2026

Python

LangChain

Neo4j

Ollama

Graph-based RAG pipeline converting unstructured text into structured triples using LLMs.
Query-aware reasoning-path retrieval over Neo4j knowledge graph with Ref(p) scoring.
Entropy-based conflict module (ΔH) comparing parametric and graph-augmented responses.
Reduced hallucination vs vector RAG on multi-hop QA benchmarks.
Integrated with Ollama for local open-source LLMs ensuring data privacy.

⬡ RajNLP-50K

2026

Python

HuggingFace

MuRIL

India's first open Rajasthani-Hindi code-switched NLP corpus of 50,000 sentences from Twitter/X & ShareChat.
Fine-tuned MuRIL models outperforming GPT-4o on sentiment analysis, NER, and toxicity detection.
Extended pipeline to feed ASR and NLU components of BolKota, a Rajasthani voice assistant.
Published dataset and model checkpoints on HuggingFace Hub.

⬡ Cardiac Edge AI

2026

Python

PyTorch

TFLite

Arduino

5-class cardiac arrhythmia detector on Arduino Nano 33 BLE via PyTorch & TensorFlow Lite.
Novel spectral knowledge-distillation loss (Lspectral) giving 140x compression (1.27M to 9.1K params, 23.4KB INT8).
∼99% F1 via multimodal ECG & PPG fusion with temporal signal preprocessing.
Real-time serial dashboard in C++ to monitor inference outputs with minimal latency.

⬡ standin

2026

Python

PyPI

pytest

LLM

A stand-in for the real LLM in your tests: record API calls once, replay them forever, fast, free, deterministic, and offline.
Hooks httpx, requests, and aiohttp, with fuzzy and semantic matchers plus built-in secret redaction.
Published on PyPI: pip install standin.

⬡ Sneakoscope

2026

Python

PyPI

Security

AI Agents

A safety hook for AI coding agents: logs every tool call and blocks file access outside the folders you allow.
Live dashboard; works with Claude Code, Gemini CLI, and Cursor.
Published on PyPI: pip install sneakoscope.

⌥ View all repos on GitHub →

eesh@os:~$ ls ./projects/ | wc -l

ACHIEVEMENTS.EXE

COMPETITIVE RATINGS (MAX)

CODEFORCES

1582

Specialist

Top 1% in India

CODECHEF

1868

4 ★ Star

Top 1% in India

LEETCODE

1967

Knight

Top 5.2% Worldwide

Problems Solved

1,500+

CONTEST HIGHLIGHTS

Global Rank 1 in CodeChef Starters 180 (Div. 3), 10,000 participants
Global Rank 75 in LeetCode Weekly Contest 446, 27,686 participants
Global Rank 78 in CodeChef Starters 183 (Div. 2)
Participated in 50+ contests across Codeforces, CodeChef & LeetCode
Kaggle Competitions Expert: silver in Kaggriculture (473 / 10,148) and bronze in Biohub cell tracking (287 / 3,947)

eesh@os:~$ cat achievements.log

SKILLS.EXE

TECHNICAL STACK

Languages

C++

Python

JavaScript

PHP

SQL

Frameworks

React.js

Node.js

Express.js

PyTorch

LangChain

Libraries / ML

NumPy

Pandas

Matplotlib

HuggingFace

TFLite

Databases

MySQL

MongoDB

Neo4j

Supabase

Firebase

Tools

Git

Linux

Postman

Arduino

Cloud / Deploy

AWS

Vercel

Netlify

Railway

RESEARCH AREAS

Reversible Data Hiding

Multiple Object Tracking

NLP / Low-Resource Languages

Edge AI / TinyML

Knowledge Graphs / RAG

Graph Neural Networks

eesh@os:~$ which skills

RESUME.EXE

RÉSUMÉ

CS Engineering at IIIT Senapati. AI/ML and NLP researcher, competitive programmer. Research interns at IIIT Vadodara and IIT Tirupati. 1500+ problems solved.

HIGHLIGHTS

Research Intern at IIIT Vadodara (Reversible Data Hiding)
Winter Research Intern at IIT Tirupati SEVA Lab (Multi-Object Tracking)
RajNLP-50K, India’s first open Rajasthani-Hindi corpus
Cardiac Edge AI: 140x compressed arrhythmia model on Arduino
Specialist on Codeforces, Knight on LeetCode
Kaggle Competitions Expert (1 silver, 1 bronze)

DOWNLOAD

↓ Download PDF ↗ Open full résumé

eesh@os:~$ cat resume.pdf