Vaishnavi Gahoi — AI Engineer & Automation Specialist

About me

I turn AI tools into real solutions

I'm a Computer Science graduate (B.Tech, RGPV 2026) who spent the last year building actual AI systems — not just tutorials. At EOXS, I designed automation workflows that replaced hours of manual supplier research with intelligent pipelines using n8n and LLMs.

My focus is on the intersection of AI automation, data analytics, and generative AI engineering. I care about one thing above all: does it work in production? Can a real team use it? Does it actually save time or surface real insights?

I'm a Google × Kaggle AI Intensive alum, a district-level chess player, and someone who genuinely gets excited about building things that didn't exist before.

♟️

District-level chess player — strategic thinking applies to system design too.

B.Tech in Computer Science

2022 – 2026

JNCT, RGPV — Rewa, MP

CGPA: 8.33 · Coursework: ML, AI, NLP, DBMS, DSA, Software Engineering

AI Generalist Intern

Aug – Oct 2025

EOXS · Remote

Built AI automation pipelines, deployed a Vapi.ai calling agent, designed KPI dashboards for internal teams

Data Analyst Intern

Jan – Feb 2025

NullClass · Remote

Cleaned & analyzed 5,000+ row datasets, created visual reports on behavioral trends and KPIs

Project-based AI Trainee

Dec 2024 – Jan 2025

TechSaksham (Microsoft & SAP) · Remote

EDA, preprocessing, ML modeling on real-world datasets in Microsoft initiative

Work experience

Where I've built things

Three internships, all remote, all building real systems.

AI Generalist Intern

Aug – Oct 2025

EOXS · Remote

Built AI automation workflows in n8n that scraped supplier websites, extracted product attributes using LLMs, and matched them to catalog entries — cutting manual qualification time by 30%
Developed an AI-driven outbound calling agent (Vapi.ai) that generated call scripts, analyzed transcripts using LLM-based scoring models, and automatically logged structured feedback into Google Sheets
Designed KPI dashboards to monitor workflow accuracy, model performance, and automation throughput for internal teams
Evaluated and deployed emerging AI tools to optimize processing speed, reduce error rates, and improve workflow reliability

Data Analyst Intern

Jan – Feb 2025

NullClass · Remote

Cleaned, transformed, and analyzed datasets (5,000+ rows) using Python and SQL to identify behavioral trends and business insights
Created visual reports and dashboards that improved clarity of patterns across categories, user segments, and performance KPIs

Project-based AI Trainee

Dec 2024 – Jan 2025

TechSaksham (Microsoft & SAP Initiative) · Remote

Developed AI-focused mini-projects involving EDA, preprocessing, and visualization across real-world datasets
Applied Python and SQL to derive insights, build models, and prepare datasets for downstream analytics tasks

Projects

Things I've actually built

Each one is a real system solving a real problem — not a tutorial copy.

🤖

Multi-Agent Career Pathfinder

A multi-agent architecture with Profile, Learning, and Career agents that generates personalized learning paths and job recommendations. Integrated search, evaluation, and reasoning agents to create a feedback loop.

↑ Personalized learning recommendations with feedback loop accuracy

PythonGemini 2.0 FlashMulti-agentAI Reasoning

GitHub ↗

🏗️

AI Data Architect

An AI system for Google's "Build the Future" showcase that converts unstructured supply chain documents into SQL-ready relational tables. Implemented entity extraction and schema mapping using a role-aligned LLM persona for hallucination-free JSON output.

↑ Zero-hallucination structured data from unstructured docs

Gemini 1.5 FlashStreamlitPythonPrompt Engineering

GitHub ↗

📊

7-Days, 7-Dashboards Power BI Challenge

Seven end-to-end dashboards built in 7 days including HR Analytics, Netflix Trends, Weather Analysis, and Sales Forecasting. Used Power Query and DAX with data modeling, measures, and forecasting logic.

↑ 7 production-ready dashboards with forecasting & KPIs

Power BIDAXExcelPower Query

HR Dashboard ↗ Sales Dashboard ↗

♿

AI Accessibility Companion

A full-stack PWA that makes digital content accessible — GPT-4o Vision for image description, Whisper STT for voice input, OpenAI TTS for audio output, and Twilio for SMS alerts. Built for users with visual or motor impairments.

↑ Vision + Voice + TTS in one deployed PWA · Real accessibility use case

FastAPIReactGroqWhisper STTTwilioRedisRailway

GitHub ↗

📰

AI News Automation Pipeline

End-to-end news aggregation system that fetches articles across 10 categories, summarizes them using Ollama (Llama 3 8B), stores them in SQLite, and delivers daily digests via Telegram Bot. Includes a Chart.js dashboard for trend visualization.

↑ 80 articles/run · 10 categories · Automated Telegram delivery

PythonOllamaLlama 3FlaskSQLiteTelegram BotChart.js

GitHub ↗

📱

Google Play Store App Analysis

Analyzed 10,000+ apps to uncover category-level performance, rating behaviors, and install trends. Used TF-IDF on 1-star reviews to extract recurring complaint themes and propose product improvements.

↑ 10,000+ apps analyzed · NLP-driven product insights

PythonPandasNLPTF-IDFMatplotlib

GitHub ↗

🛍️

Shopping Trends Analysis

Deep analysis of consumer shopping patterns using Python and CSV datasets. Surfaced behavioral trends, seasonal patterns, and category insights to support data-driven business decisions.

↑ Consumer pattern insights for business strategy

PythonPandasMatplotlibEDA

GitHub ↗

🏠

Home Decor Market Analysis

Analyzed market trends in the home decor and lifestyle sector, identified top competitors, and evaluated their market strategies. Provides insights into pricing trends and customer preferences.

↑ Competitor strategy + pricing intelligence insights

PythonMarket AnalysisCompetitor Research

GitHub ↗

📝

Changelog AI

Automated changelog generator that converts raw Git commits into clean, human-readable release notes using LLMs. Deployed and used by real teams to skip the manual changelog writing step entirely.

↑ Eliminates manual changelog writing · Live deployment

PythonLLMGitAutomationFastAPI

GitHub ↗ Live ↗

🔁

Self-Healing RAG Pipeline

A retrieval-augmented generation system built with LangGraph that detects low-confidence retrievals and automatically re-routes or reformulates queries. Reduces hallucinations without human intervention.

↑ Auto-corrects retrieval failures · Zero human retry needed

LangGraphRAGPythonLLMAgentic

GitHub ↗ Live ↗

✉️

AI Cold Outreach Personalizer

Generates hyper-personalized cold emails by scraping target company data and feeding it into an LLM pipeline. Each email references real context about the recipient — not generic templates.

↑ Context-aware personalization at scale · Real company data

PythonLLMWeb ScrapingAutomationPrompt Engineering

GitHub ↗ Live ↗

📈

India Employment Pulse — EPFO Dashboard

Live analytics dashboard tracking India's formal employment trends using EPFO payroll data. Visualizes monthly net subscriber additions by age group, state, and industry — making public data actually readable.

↑ Real EPFO data · Public deployment · India employment trends

PythonStreamlitPlotlyEPFO DataData Viz

GitHub ↗ Live ↗

✂️

Cutsy — AI Video Trimmer

A free, open-source alternative to paid AI video editors. Upload a video, describe the cut in plain English ("give me the best 30 seconds"), and Whisper transcribes it, Llama 3.1 picks the segment, and FFmpeg cuts it — no timeline needed.

↑ Fully free stack · No paid API required

PythonWhisperGroq / Llama 3.1FFmpeg

GitHub ↗ Live ↗

🔗

n8n Automation Portfolio

A collection of production-ready n8n workflows built during my EOXS internship: AI news aggregation, a Vapi.ai voice-agent webhook handler, LinkedIn persona classification, X/Twitter scraping, and automated lead email verification (Apollo, Emailable).

↑ 50+ domains processed · 30% less manual data alignment

n8nVapi.aiOpenAIApollo APIWebhooks

GitHub ↗

🕵️

DataScout — AI Data Analyst

A BI platform that lets anyone query CSV datasets in plain English. Converts questions to SQL with Gemini 1.5 Flash, auto-generates visualizations, profiles data health, and produces boilerplate code in SQL, Pandas, and C++.

↑ Natural language → SQL, charts & multi-language code

StreamlitGemini 1.5 FlashPandasSQLite

GitHub ↗ Live ↗

🔍

AI Job Analyser

Paste any job description and get a legitimacy score with red/green flags on one side, and 10 tailored interview questions with ideal answer structures and skills to highlight on the other. Built as Day 5 of a 7-day AI build series.

↑ Instant legitimacy scoring · Tailored interview prep

StreamlitGroq / Llama 3.1Python

GitHub ↗ Live ↗

🧹

AI Data Quality Fixer

Upload any messy CSV or Excel file — AI scans for nulls, duplicates, outliers, inconsistent casing, and mistyped numeric columns, then lets you pick which fixes to apply before downloading the cleaned data. Day 6 of a 7-day build series.

↑ One-click detection + fixing of 7 common data issues

PythonPandasStreamlitGroq API

GitHub ↗ Live ↗

📝

AI Meeting Notes Fixer

Turns rough, unstructured meeting notes into clean, organized summaries with action items — part of the same 7-day AI build series as the Job Analyser and Data Quality Fixer.

↑ Messy notes → structured, actionable summaries

PythonLLMStreamlit

GitHub ↗ Live ↗

🎓

Smart ATS Resume Optimizer

Scores a resume against a job description (0-100%), flags missing keywords, rewrites weak bullet points, and predicts 5 likely interview questions — built as a personal AI career coach beating Applicant Tracking Systems.

↑ Match scoring, keyword gaps & interview prediction in one tool

StreamlitGemini 2.0 FlashPyPDF2

GitHub ↗ Live ↗

Vaishnavi
Gahoi

I turn AI tools into real solutions

Where I've built things

Things I've actually built

What I work with

Credentials

Let's build something together

VaishnaviGahoi

I turn AI tools into real solutions

Where I've built things

Things I've actually built

What I work with

Credentials

Let's build something together

Vaishnavi
Gahoi