STATUS: OPEN TO SDE & BACKEND ROLES (2026/2027)

// welcome.md — Software Development Engineer

/**

* @author Suhani Singhania

* @focus Distributed Data Systems, Backend APIs, NLP & Cloud Architecture

* @experience American Express (SDE Intern) & Adobe Systems (SDE Intern)

*/

const engineer = {

Suhani
Singhania▌

role: "SDE — Distributed Systems & Backend"

B.Tech CSE undergrad at VIT Vellore. SDE Intern at American Express (optimizing Apache Spark pipelines over 3M+ records) and Adobe Systems (building semantic document search cutting latency from 800ms to <120ms).

3M+ Records Pipeline

<120ms Search Latency

94.8% AI Detection Acc

3 SDE Internships

🎮 Play Load Balancer Game ⚡ Launch System Simulators 📂 View Shipped Code

};

// 01 — about.md

Reads stack traces before coffee.

I'm a Computer Science undergrad who cares deeply about real-world software performance when high-throughput traffic meets distributed infrastructure.

Most of my engineering work sits at the intersection of distributed data pipelines and high-concurrency backend microservices — tuning Apache Spark & Hadoop jobs handling millions of records, building Redis-backed distributed rate limiters, cutting document search latency from 800ms down to <120ms with PyTorch vectors, and co-authoring NLP research.

🚀 Spark & Distributed Data

⚡ Redis & Rate Limiting

🧠 NLP & PyTorch Embeddings

☁️ AWS Serverless Architecture

period: "2023 — 2027 (expected)"

Vellore Institute of Technology

B.Tech, Computer Science & Engineering — Vellore, India

Core: DSA, Operating Systems, DBMS, Cloud Computing, Computer Networks

period: "2021"

Kendriya Vidyalaya

Senior Secondary (CBSE) — Kathmandu, Nepal

Stream: Physics, Chemistry, Mathematics, Computer Science

// 02 — experience.js

export const experience = [

AmEx

{

role: "Software Development Engineer Intern",

company: "American Express (India) Pvt. Ltd.",

location: "Gurugram, HR",

period: "May 2026 — Present", CURRENT ROLE

stack: ["Apache Spark", "Hadoop MapReduce", "Apache Pig", "Java", "PySpark"],

highlights: [

Building and optimizing a distributed data processing layer on Apache Spark, Hadoop MapReduce, and Apache Pig, processing over 3 Million transaction records.
Engineered custom partition tuning, broadcast joins, and lazy-evaluation strategies, slashing pipeline execution time from 47 min to under 15 min (68% speedup).
Collaborating with a team of 8 senior SDEs, data scientists, and product managers in 2-week Agile sprints with rigorous code reviews.

]

},

Adobe

{

role: "Software Development Engineer Intern",

company: "Adobe Systems India Pvt. Ltd.",

location: "Noida, UP",

period: "May 2025 — Jul 2025",

stack: ["Python", "PyTorch", "Elasticsearch", "REST APIs", "Docker", "Kubernetes", "AWS"],

highlights: [

Delivered a production-ready microservice feature for Acrobat's Fill & Sign engineering team serving global users.
Architected REST APIs with unit/integration testing, containerized with Docker, and deployed via Kubernetes on AWS CI/CD pipelines.
Engineered a semantic document search engine pairing PyTorch sentence embeddings with Elasticsearch, cutting search query latency from 800ms to <120ms across 2M+ document assets.

]

},

SentiAid

{

role: "Web Development Intern",

company: "SentiAid",

location: "Remote",

period: "May 2025 — Jul 2025",

stack: ["MongoDB", "Express.js", "Node.js", "React"],

highlights: [

Developed healthcare analytics modules using MongoDB, Express.js, and Node.js — improving data retrieval speed by 25% for daily practitioners.
Optimized database index strategies and async controller calls, reducing dashboard render time from 4s to 2.5s and enabling a 20% boost in concurrent users.

]

}

];

// 03 — projects.json

Production Code & Systems Shipped.

"careerverse": FEATURED FULL-STACK

CareerVerse — AI Intelligence Platform

GitHub ↗

Full-stack career intelligence platform with AI-driven resume analysis and personalized job recommendations. Built with JWT authentication, MongoDB data layer, and Google Gemini API integration for real-time document scoring and automated SMTP alert notifications.

FastAPIReactViteMongoDBGemini APIJWT

"billstream": DISTRIBUTED INFRA

BillStream — Multi-Tenant Billing & Rate Limiter

GitHub ↗

Multi-tenant API billing and distributed rate-limiting platform featuring Redis-backed Token Bucket and Sliding Window algorithms, JWT authentication, and clean layered microservices architecture.

JavaSpring BootPostgreSQLRedisDocker

"trendsage_ai": ANALYTICS ENGINE

TrendSage AI — Wellness Signal Analytics

GitHub ↗

AI-powered wellness intelligence platform capturing and analyzing 150+ D2C market signals in real time with interactive timeline visualizations and predictive trend models.

React.jsTypeScriptNode.jsExpress.jsChart.js

"email_overload_organizer": NLP & ML PIPELINE

Email Overload Organizer — NLP spaCy Triage

GitHub ↗

NLP pipeline scoring email importance with SentenceTransformer + Stacking Classifier (Logistic Regression, Random Forest, Linear SVM), followed by spaCy NER extracting people, organizations, deadlines, and action items. Trained on 20K+ emails at ~80% accuracy.

PythonSentenceTransformersscikit-learnspaCyStreamlit

// 04 — simulators.json

Interactive SDE System Simulators.

Real-time interactive sandboxes demonstrating rate limiting algorithms, NLP entity extraction, ATS resume scoring, and vector search latency.

SIMULATOR 01

BillStream — Distributed Token Bucket Rate Limiter

● REDIS ACTIVE

Simulate Redis token bucket traffic regulation under heavy concurrent request surges. Adjust request rate and capacity to see live HTTP 200 (Allowed) vs HTTP 429 (Rate Limited) responses.

Requests / Min (RPM): 120

Bucket Capacity: 10

REDIS TOKEN BUCKET (10 / 10 TOKENS)

[SYSTEM] Redis Rate Limiter initialized. Token refill rate: 2 tokens/sec.

SIMULATOR 02

Adobe Acrobat — Semantic Vector Search Latency Visualizer

● PYTORCH EMBEDDINGS

Benchmarking linear document keyword search vs PyTorch 768-dim sentence embedding cosine similarity across 2,000,000 document assets.

Baseline Keyword Search

800 ms

Linear string scan (O(N))

PyTorch Vector Search (Suhani's Opt)

118 ms

6.77x Faster (Cosine Similarity + Index)

[RESULTS] Matched "Acrobat_Enterprise_Security_Spec.pdf" — Similarity: 0.942

SIMULATOR 03

CareerVerse — AI Resume Keyword & ATS Matcher

● GEMINI API ENGINE

Test real-time AI scoring of resume text against target software engineering job descriptions.

Select Target Role:

94%

ATS MATCH SCORE

Distributed Systems / Spark:

Backend REST & DBs (Redis/Mongo):

Cloud DevOps & Docker:

SIMULATOR 04

Email Overload Organizer — spaCy NER & Priority Scoring

● SPACY & SCIKIT-LEARN

Runs spaCy Named Entity Recognition (NER) and stacking classifier to triage inbox messages in real time.

Select Sample Email:

IMPORTANCE SCORE: 98 / 100 (CRITICAL)

ORGANIZATIONS: American Express

SYSTEM CONCERNS: Spark Cluster, Memory Leak

RECOMMENDED ACTION: Trigger PagerDuty alert & re-route cluster partition traffic.

// 05 — games.js

🎮 SDE DEV ARCADE & INTERACTIVE GAMES

Because building great software is engaging & interactive.

Prove your engineering skills, balance real-time traffic loads, smash production bugs, or challenge the SDE knowledge matrix.

SCORE 0

TRAFFIC SLA 100%

ACTIVE WAVE 1

CLUSTER STATUS HEALTHY

⚡ DISTRIBUTED TRAFFIC LOAD BALANCER

High-throughput HTTP requests are streaming into the AmEx cluster! Route packets to Spark worker nodes (Node A, B, C) to keep CPU under 100% and prevent 500 server errors.

Controls: Click/Tap Worker Nodes OR press keys [1] [2] [3] to balance traffic!

BUGS FIXED 0

COMBO x1

SERVER HEALTH 100%

👾 PRODUCTION BUG SMASHER ARCADE

Production bugs (NullPointer, Memory Leak, Deadlock, Race Condition) are dropping into main branch! Click or tap bugs rapidly to hotfix before server health reaches 0%.

QUESTION 1 / 5 SCORE: 0

What Spark optimization technique avoids expensive shuffles by copying small data frames across all worker nodes?

// 06 — research.md

Co-Authored NLP & Cloud Research.

type: "research paper" 94.8% ACCURACY

Cloud-Based AI Plagiarism Detection System Using NLP and Machine Learning

Authors: Suhani Singhania, Arushi Bhakuni, Shubhangi Goyal, Pavithra B

Engineered a hybrid detection pipeline pairing a fine-tuned RoBERTa transformer for AI-generated text synthesis with statistical plagiarism algorithms (Jaccard trigram + TF-IDF cosine).
Architected a serverless AWS cloud infrastructure (AWS Lambda, SQS, DynamoDB) executing async AI inference alongside instant synchronous plagiarism lookups.
Achieved 94.8% AI-detection accuracy, outperforming conventional n-gram baselines while sustaining sub-2-second total processing time.

RoBERTaNLPAWS LambdaSQSHuggingFaceDynamoDB

type: "literature survey" SURVEY & TAXONOMY

AI-Powered Cloud Resource Management: Self-Healing and Predictive Infrastructure

Authors: Suhani Singhania, Ananya, Samhitaa, Swatika

Synthesized 25+ peer-reviewed papers on AI-driven cloud optimization into a comprehensive taxonomy spanning supervised learning, deep neural nets, and reinforcement learning.
Analyzed self-healing cloud architectures cutting operational downtime by 85% and LSTM workload forecasting achieving 95% predictive accuracy.
Evaluated AI-driven threat mitigation systems operating at 97% detection accuracy with under 5% false positive rates.

Literature SurveySelf-Healing InfrastructureCloud AILSTM Forecasting

# 07 — skills.yaml

The Engineering Toolbelt.

languages:

Java (Spring Boot / OOP)

Python (PySpark / PyTorch / NLP)

JavaScript / TypeScript

C++ & SQL

frameworks_libs:

Spring BootFastAPIReact.jsNode.jsExpress.jsPyTorchscikit-learnspaCyStreamlit

databases_caching:

PostgreSQLMongoDBRedis (Rate Limiting)Elasticsearch

distributed_big_data:

Apache SparkHadoop MapReduceApache PigBroadcast JoinsPartition Tuning

cloud_devops:

AWS LambdaSQSAPI GatewayS3DynamoDBDockerKubernetesGitGitHub Actions

core_competencies:

REST APIsJWT AuthDistributed SystemsData Structures & AlgorithmsNLP

// 08 — leadership.md

Engineering Leadership & Initiatives.

Outreach Lead

IEEE Women in Engineering (WIE), VIT

2024

Spearheaded community outreach initiatives, tech workshops, and hackathons empowering female engineering students.

Media & Finance Manager

RoverX VIT

2024

Managed financial budgeting, sponsorship acquisition, and public media for VIT's flagship University Rover Challenge team.

Campus Ambassador

Techfest, IIT Bombay

Jul 2025 — Dec 2025

Represented Asia's largest science and technology festival across VIT, driving student hackathon participation.

# 09 — contact.sh

Let's build
something that scales.

Actively seeking SDE, Backend, and Distributed Systems roles starting 2026/2027.

$ echo $EMAIL

suhanee14@gmail.com

$ echo $PHONE

+91-9380825153

$ open --social

linkedin ↗ github ↗

$ ▌

✉ Send Email (suhanee14@gmail.com)

SuhaniSinghania▌

Reads stack traces before coffee.

Vellore Institute of Technology

Kendriya Vidyalaya

export const experience = [

Production Code & Systems Shipped.

CareerVerse — AI Intelligence Platform

BillStream — Multi-Tenant Billing & Rate Limiter

TrendSage AI — Wellness Signal Analytics

Email Overload Organizer — NLP spaCy Triage

Interactive SDE System Simulators.

BillStream — Distributed Token Bucket Rate Limiter

Adobe Acrobat — Semantic Vector Search Latency Visualizer

CareerVerse — AI Resume Keyword & ATS Matcher

Email Overload Organizer — spaCy NER & Priority Scoring

Because building great software is engaging & interactive.

⚡ DISTRIBUTED TRAFFIC LOAD BALANCER

👾 PRODUCTION BUG SMASHER ARCADE

What Spark optimization technique avoids expensive shuffles by copying small data frames across all worker nodes?

Co-Authored NLP & Cloud Research.

Cloud-Based AI Plagiarism Detection System Using NLP and Machine Learning

AI-Powered Cloud Resource Management: Self-Healing and Predictive Infrastructure

The Engineering Toolbelt.

Engineering Leadership & Initiatives.

Outreach Lead

Media & Finance Manager

Campus Ambassador

Let's buildsomething that scales.

Suhani
Singhania▌

Let's build
something that scales.