AI / ML · Full-Stack · Research

Peesari Sathvik Reddy

|

AI/ML engineer in training focused on machine learning, distributed systems, and full-stack development - with experience spanning research, product development, and real-world deployments.

Currently B.Tech AI/ML, CBIT
Batch 2023 - 2027
Location Hyderabad, India

Who I Am

I am an Artificial Intelligence & Machine Learning undergraduate at Chaitanya Bharathi Institute of Technology (CBIT) with a strong interest in building intelligent systems, full-stack applications, and research-driven solutions that solve real-world problems.

My journey in technology began with curiosity and has evolved into a passion for transforming ideas into impactful products. Over the years, I have worked on projects spanning machine learning, computer vision, natural language processing, and web development.

"Bide your time. Hide your strength."

To me, it is a reminder to focus on preparation, continuous improvement, and letting results speak louder than words. Beyond academics, I lead teams and organize technical initiatives as Vice President of the Robotics & Innovation Club at CBIT.

When I'm not building or researching, I write on Medium - stories about technology, health, and the quieter parts of personal growth.

Academic Journey

2009 - 2021

St. Paul's High School

Secondary School Certificate (SSC)

CGPA: 10

Hyderabad, India

2021 - 2023

Sri Chaitanya Junior College

Intermediate Education (MPC)

981 / 1000

Hyderabad, India

2023 - 2027 Current

Chaitanya Bharathi Institute of Technology

B.Tech - Artificial Intelligence & Machine Learning

CGPA: 9.20

Hyderabad, India

Technical Skills

Languages

Python JavaScript SQL HTML CSS

Frameworks & Libraries

React.js Node.js Express.js Flask TensorFlow Scikit-learn

Machine Learning

Machine Learning Deep Learning Computer Vision NLP MediaPipe PyTorch

Tools & Technologies

MongoDB BeautifulSoup Git / GitHub Linux Power BI Supabase

Things I've Built

AIML

Zenvia Fashion AI System

Intelligent fashion platform with size estimation using MediaPipe Pose, color analysis with ResNet50, virtual wardrobe, and weather-based outfit suggestions.

Paper presented at ICAIATI-2025 (under publication)

PythonFlaskMLMediaPipe
Click to flip

Zenvia - Key Highlights

  • MediaPipe Pose for body measurement estimation
  • ResNet50 for color palette analysis
  • Real-time weather API for outfit suggestions
  • Virtual wardrobe with outfit history
  • Research published at ICAIATI-2025
AIML

RAY - Resume-based Application Yield

A resume-based job opportunity discovery platform that matches resumes to real job openings using NLP-based relevance scoring. Features job freshness indicators, skill insights, and saved jobs.

PythonStreamlitNLPscikit-learn
Click to flip

RAY - Key Highlights

  • NLP-powered resume-to-job matching
  • Freshness indicator for job postings
  • Skill gap analysis per listing
  • Save & track applied jobs
  • Deployed on Streamlit Cloud
AIML

PathVQA - Medical VQA

Multimodal VQA for pathology images using CNN encoders, attention mechanisms, and bilinear fusion - achieved 60.39% exact-match accuracy.

PyTorchCNNsAttentionMultimodal AI
Click to flip

PathVQA - Key Highlights

  • CNN encoder for pathology image features
  • Attention mechanisms for focus regions
  • Bilinear fusion of visual + text modalities
  • 60.39% exact-match accuracy on PathVQA dataset
  • Built with PyTorch end-to-end
SDE

SETHU - Campus Management Platform

Role-based campus management for CBIT - timetable, exams, almanac, AI study planner, requests/approvals, notifications. Built with Next.js, Supabase, FastAPI.

Next.jsSupabaseFastAPIMistral AI
Click to flip

SETHU - Highlights

  • 4 roles: Student, Faculty, HOD, Admin
  • Timetable / exam / almanac with PDF import (OCR + Mistral)
  • AI study planner based on schedule & annotations
  • Request workflows with auto-generated PDFs
  • Row-Level Security + audit log
SDE

CONTINUUM - Healthcare Continuity

Offline-first healthcare platform - patients maintain a portable health record, share with doctors, async consultations, live video calls, and automatic sync.

ReactNode.jsMongoDBWebRTCDexie
Click to flip

CONTINUUM - Highlights

  • Health timeline - conditions, meds, consultations
  • Async voice-based consultations
  • Live WebRTC video calls with recording
  • Offline-first (IndexedDB) with automatic sync
  • Doctor dashboards & analytics
AIML

Hindi News Classification

End-to-end system to classify Hindi news headlines using fine-tuned IndicBERTv2. Features OCR, web scraping, and a Flask web interface.

FlaskIndicBERTEasyOCRBeautifulSoup
Click to flip

Hindi News - Highlights

  • Fine-tuned IndicBERTv2 (79.57% accuracy)
  • Live scraping from 4 major Hindi news sites
  • OCR extraction from images/PDFs
  • Confidence-based predictions
  • Deployed on Hugging Face Spaces
Data Science

Telecom Churn Analysis

Predicted customer churn for 7,043 telecom customers using RandomForest. Optimized decision threshold for business ROI (₹161,500) rather than accuracy.

PythonRandomForestscikit-learnPandas
Click to flip

Telecom Churn - Highlights

  • 82.4% recall, 861 at-risk customers identified
  • Threshold tuned for ROI (0.45 vs default 0.5)
  • Top churn drivers: new customers, tenure, tech support
  • Net business value: ₹161,500
  • Actionable risk segments with budget
Data Science

Olist E-Commerce Analytics

End-to-end analytics pipeline on 100k+ Brazilian orders. SQL queries and a Power BI executive dashboard answering key business questions.

SQLMySQLPower BIPython
Click to flip

Olist - Highlights

  • 97% orders delivered successfully
  • Top categories: Health & Beauty, Watches
  • Late deliveries drop ratings from 4.29 to 1.86
  • Repeat purchase rate only 3%
  • Power BI dashboard with 3 pages
SDE

Syncpad - Collaborative Interview Platform

Real-time collaborative coding platform with live editing, audio/video, quiz system, and resume analyzer. Built in 24h hackathon - Special Mention among 500 teams.

ReactNodeLiveblocksWebRTCMongoDB
Click to flip

Syncpad - Highlights

  • Collaborative code editor (Liveblocks + Monaco)
  • Audio/video via LiveKit WebRTC
  • Quiz module & daily gamified quiz
  • Resume analyzer with Mistral AI
  • Special Mention - CBIT Hacktoberfest 2025
AIML

EngiQuery - AI Document Agent

AI agent for technical document analysis - upload manuals, ask questions, compare specs, detect contradictions, and email reports. Uses Groq LLM + ChromaDB.

FastAPIGroqChromaDBsentence-transformers
Click to flip

EngiQuery - Highlights

  • Multi-document Q&A with citations
  • Cross-document comparison & contradiction detection
  • Compliance gap checking
  • Agent loop with tool selection
  • Email reports via Gmail

On Medium

Beyond code and research, I write about technology, health, and the quieter parts of personal growth.

Credentials

AWS Certified Cloud Practitioner

Amazon Web Services

Salesforce Certified Agentforce Specialist

Salesforce · 2025

Beyond the Keyboard

Vice President

Robotics & Innovation Club, CBIT

Jun 2024 - Feb 2026

3rd Place - HackWeek 2026

Secured 3rd place individually among 1,200+ participants in a COSC week-long hackathon.

Week-long hackathon

Special Mention

CBIT Hacktoberfest 2025 - recognized among 500+ teams for Syncpad, a collaborative coding interview platform.

24-hour virtual hackathon

NCC Junior Cadet

1(T) Naval Unit, Hyderabad

2018 - 2022

Let's Connect

Actively seeking internship opportunities in AI/ML and software development. Open to collaborations, research projects, and anything interesting.

reddysathvik2005@gmail.com Copied!

Hyderabad, India