Open to full-time roles·IST

Manvi Mittal Full stack engineer.

I build web apps, REST APIs and machine-learning services with React, Next.js, TypeScript, Node.js, Python and SQL — currently shipping calibration and inspection platforms at Ezentech India.

  • 0%fewer missed calibration schedules
  • 0%lower end-to-end processing time
  • 86.4%test accuracy on my ML classifier
Portrait of Manvi Mittal
manvi.jpg · Saharanpur, IN
whoami.json
{
  "name": "Manvi Mittal",
  "role": "Full Stack Engineer",
  "now": "Ezentech India",
  "stack": ["Next.js", "TS",
            "Node", "Python"],
  "open_to_work": true
}
ReactNext.jsTypeScriptNode.jsExpressPythonFastAPIscikit-learnJavaMySQLMongoDBDockerAzureGemini APIVercel ReactNext.jsTypeScriptNode.jsExpressPythonFastAPIscikit-learnJavaMySQLMongoDBDockerAzureGemini APIVercel
Manvi Mittal
hello, world 👋

01 — about

Hi, I'm Manvi.

I'm a full stack software engineer with a year of experience building production web applications and REST APIs.

At Ezentech India I build the calibration monitoring and digital inspection platforms that replaced spreadsheets and paper checks for operations teams. On my own time I build end-to-end projects — an explainable ML classifier served with FastAPI, a Gemini-powered assistant, and a production Next.js site.

I care about developer experience, performance and code that's easy to maintain.

based in
Saharanpur, UP, India
currently
Software Engineer @ Ezentech
education
B.E. CSE · Chitkara University
focus
Full stack · APIs · applied ML

02 — selected work

Things I've built.

Four projects, each shown with real screenshots, real code or real numbers.

01machine learning

NewsLens

An explainable news-topic classifier. Paste an article and it returns the topic, ranked probabilities and the words that drove the decision — served by a production-style FastAPI backend.

test accuracy
0.4%
macro-F1
0.863
held-out docs
2,282
topics
6
127.0.0.1:8000
NewsLens app classifying a space article Prediction with probabilities and keyword explanation Auto-generated Swagger API docs NewsLens in light mode

model selection · 5-fold CV macro-F1

  • Logistic Regression0.864
  • Linear SVM0.872
  • Complement NB ★0.877

winner tuned with GridSearchCV (α = 0.3) · bars scaled 0.78 → 0.92

per-class F1 · test set

  • Baseball0.91
  • Automobiles0.87
  • Space0.87
  • Medicine0.86
  • Graphics0.83
  • Politics0.82
app/ml.py — keyword explainer
def _keywords(self, row, class_idx: int, top_n: int) -> list[dict]:
    """Terms in the input that pushed the prediction towards the class."""
    if self._weights is None or row.nnz == 0:
        return []
    contrib = row.data * self._weights[class_idx, row.indices]
    order = np.argsort(contrib)[::-1][:top_n]
    return [{"term": str(self.vocabulary[row.indices[i]]),
             "weight": round(float(contrib[i]), 4)}
            for i in order if contrib[i] > 0]

rest api · /api/v1

  • GET/health
  • GET/model
  • POST/predict
  • POST/predict/batch
  • GET/history
  • DEL/history
  • GET/stats
Python 3.12scikit-learnTF-IDFFastAPIPydanticSQLitepytestDockerGitHub Actions CI

Batch inference · prediction history & stats · non-root Docker image with healthcheck · tests run without downloading the dataset.

02source on GitHub

AI Legal Assistant

A full-stack chatbot on the Google Gemini API — multi-turn chats with saved history, a prompt library and paid plans.

View source ↗
ai-project ⎇ main
TSgenAi.ts
import { Content, GoogleGenerativeAI } from "@google/generative-ai"

const GENAI_MODEL_NAME = 'gemini-2.0-flash';

export function generateTextResponse(historyConversation: Content[]) {
  return new Promise(async (resolve, reject) => {
    try {
      const genAI = new GoogleGenerativeAI(GENAI_API_KEY);
      const model = genAI.getGenerativeModel({
        model: GENAI_MODEL_NAME,
        systemInstruction: { role: "system", parts: [...] }
      });
      const result = await model.generateContent({
        contents: historyConversation
      })
      return resolve(result.response.text());
    } catch (err) {
      return reject(GENAI_ERROR);
    }
  })
}
src/utils/server/genAi.tsfrom the real repo

stack

Next.js 15React 19TypeScriptGemini 2.0 FlashMongoDBMongooseTailwindFramer Motion

architecture

React chat UI /api/aimodel/gentext Gemini 2.0 Flash
MongoDBhistory · limits WorkOSauth Razorpayplans Resendemail
20/day

Prompt limit per user, reset daily and enforced on the server.

PDF+ Word

Answers export to documents for sharing.

12

SEO landing pages for legal practice areas, with sitemaps.

↺

Full chat history sent each turn — chats can be saved, renamed, deleted.

03Ezentech India · Sep 2025 — now

Calibration & Inspection Platforms

role
Software Engineer
scope
Frontend · backend · APIs
data
SQL · REST services
users
Operations & quality teams
missed calibration schedules 0%↓
end-to-end processing time 0%↓
record accuracy 0%↑
beforeafter — what I built
Spreadsheet trackingREST API driven services as the single source of records
Missed calibration schedulesCalibration monitoring app with automated, system-generated alerts
Paper-based quality checksDigital inspection system with standardized capture across inspection points
Manual data entryRemoved — data captured digitally at each inspection point
Slow issue resolutionReal-time dashboards with automated notifications

Internal tools — screenshots are confidential. Figures from production use.

ramaya-ten.vercel.app
Ramaya Grand home page Venues page Gastronomy page
04live

Ramaya Grand

Multi-page Next.js site for a venue & events brand — built solo from UI to deployment.

  • 14 routes, event inquiry form, hero video and galleries.
  • Continuous deployment from GitHub to Vercel.
Next.jsReactTypeScriptVercel

03 — experience

Where I've worked.

Full résumé (PDF) ↓
  1. Sep 2025present

    Software Engineer

    Ezentech India Pvt. Ltd.Saharanpur, IN
    • Designed, built and shipped a calibration monitoring web app with automated alerts — 60% fewer missed calibration schedules.
    • Replaced manual spreadsheet tracking with REST API services — 40% better record accuracy, 55% lower processing time.
    • Built a digital inspection system that replaced paper-based quality checks.
    • Real-time dashboards with automated notifications for operations teams.
    • Code reviews, debugging and production issue resolution across frontend and backend; tests and docs for shipped features.
  2. Jun 2025Aug 2025

    Software Engineer

    Cloud InfotechIndia
    • Implemented and tested features across distributed storage, indexing and querying components — improved query response time.
    • Turned ambiguous problem statements into scoped technical designs; extended existing services and optimized query paths.
    • Shipped reviewed, tested code every two-week Agile sprint with product and engineering teams.

languages

JavaScript · TypeScript · Python · Java · SQL · HTML5 · CSS3

frontend

React · Next.js · React Spectrum · Tailwind CSS · accessible, responsive UI

backend & apis

Node.js · Express.js · FastAPI · REST API design · authentication

databases

MySQL · MongoDB · SQLite · schema design · indexing · query optimization

machine learning

scikit-learn · TF-IDF · model selection & CV · explainability · Gemini API

cloud & deploy

Azure · Docker · Vercel · Google Cloud fundamentals

testing & ci

pytest · unit testing · GitHub Actions · Git-based CI/CD

practices

Agile / Scrum · code review · debugging · documentation

04 — education

Where I learned.

Four years of computer science, plus the courses and trainings I took along the way.

bachelor of engineering · 2021 — 2025

Chitkara University

B.E. in Computer Science & Engineering

Himachal Pradesh, India

  1. 2021started B.E.
  2. 2022DBMS cert
  3. 2023JS · SE certs
  4. 2025graduated

core coursework

01Data Structures & Algorithms 02Database Management Systems 03Operating Systems 04Computer Networks 05Object Oriented Programming

during university

  • Web Developer Intern — Bharat Intern
  • Student Trainee — Google Cloud (cloud fundamentals)

school

Ram Krishna Mehta Inter College

Senior Secondary & High School · 2018 — 2021

certifications

  • JavaScriptApr 2023
  • Software EngineeringFeb 2023
  • Database Management SystemsSep 2022
  • Introduction to CybersecurityApr 2023
  • Introduction to Blockchain TechnologiesJan 2023

05 — pinboard

Screens & notes.

Real screens from my projects, plus the toolbox behind them. Click to enlarge, ★ to save.

06 — ask-ai

Ask about me.

Short on time? Ask anything about my skills, projects or availability. Answers come straight from my résumé.

✦ ask-manvi ● ready
›

07 — contact

Let's build
something good.

✦ask ai