Software Engineer • Mumbai, India
Open to Full-Time SDE / Backend Roles

Vedant
Deshmukh

Software Engineer specializing in backend systems, distributed architectures, and AI-powered applications.

My experience spans software engineering, backend development, full-stack systems, and AI research. I enjoy building scalable APIs, distributed services, and developer tools, and I'm currently exploring applied AI, intelligent agents, and LLM-powered systems.

View Resume

Experience

Growth Infi

Software Engineer (AI Automation)

Mar 2026 – Present

Growth Infi

  • ▹Engineered scalable backend APIs and serverless services using Next.js, TypeScript, and Supabase to power AI-driven workflow automation across 5+ production workflows.
  • ▹Built a domain enrichment pipeline that transformed 50K+ company names into validated, ranked company domains using search, validation, and scoring, reducing manual prospecting time by 95%.
  • ▹Designed and implemented distributed background job processing using BullMQ and worker services to orchestrate Gmail campaigns and calendar invitations with retries, exponential backoff, rate limiting, and fault-tolerant execution.
  • ▹Developed production-grade email verification pipelines integrating third-party services and custom validation heuristics, processing 5K+ email addresses while reducing verification costs by 70% through intelligent validation and caching.
  • ▹Engineered an automated data collection pipeline that analyzed Amazon Sponsored Products across 100+ keywords, extracted brand metadata, and generated appearance-based analytics for competitive market research.
DeepCytes CyberLabs (UK)

AI Intern

Jun 2024 – Dec 2024

DeepCytes CyberLabs (UK)

  • ▹Developed deepfake detection systems using CNN–LSTM pipelines with high accuracy.
  • ▹Built real-time inference pipeline under 500ms latency for live video analysis.
  • ▹Optimized preprocessing pipelines by 30% through vectorization and efficient data loading.
IIT Bombay

Full Stack Intern

Jan 2024 – Jul 2024

IIT Bombay

  • ▹Engineered a benchmarking platform for Indic language input using PHP and MySQL, handling complex session-to-phrase mapping for 1,000+ concurrent users.
  • ▹Designed a research orchestrator allowing for 'Between-Subject' and 'Within-Subject' study configurations with secure, unique URL-based user tracking.
  • ▹Integrated real-time analytics using Chart.js to visualize CPM, WPM, and error trends, automating the data pipeline and reducing manual research computation by 80%.

Tech Stack

Languages

TypeScriptJavaScriptJavaC++PythonSQLPHP

Frameworks & Backend

Node.jsNext.jsExpressFastAPIFlaskReact

Databases & Arch

PostgreSQLMySQLMongoDBRedisKafka

Cloud & Devops

DockerKubernetesAWSCI/CDGitPostman

AI Engineering

RAGAI AgentsPyTorchTensorFlowScikit-learnNumPyPandas

Backend Systems

MicroservicesREST APIsSystem DesignAuthenticationRate LimitingScalabilityUnit Testing

Projects

FynVid

FynVid

Live

Architected a scalable, high-throughput video streaming platform, leveraging Apache Kafka for seamless, asynchronous event processing and Redis for robust rate limiting. Optimized performance and reliability with an NGINX load-balancing tier.

KafkaRedisDockerMongoDBJavascriptExpressReactNGINX
FoodShare Web

FoodShare Web

Live

Developed a comprehensive full-stack platform enabling food enthusiasts to discover, share, and manage recipes. Features include secure JWT authentication, rich recipe management, and an integrated AI-powered 'Chefbot' for a personalized user experience.

DockerMongoDBJavascriptExpressReact
IRS Jamming Detection

IRS Jamming Detection

Engineered an ML-based detection framework, utilizing a hybrid CNN-LSTM model, to safeguard UAV base stations from jamming attacks. Validated performance with ns-3 simulations, achieving a detection accuracy of 87.26%.

ns-3mmwave OPEN SOURCEC++PyTorchLSTM

Application to record Text-Input Studies

Developed and deployed a specialized research platform, facilitating over 1000 data collection sessions for Indian language typing studies. Streamlined the process with automated data pipelines and a secure, efficient backend.

PHPMySQLJavaScript

Publications

Nov 2024 • Springer Nature Switzerland

IDCText: An Application for Conducting Text Input Research Studies in Indian Languages

Vedant Deshmukh, Anirudha Joshi

Released a web-based research tool for text entry studies in Indian languages. Automatically calculates WPM, CER, and TER metrics.

Apr 2024 • Springer Nature Switzerland

Skin Cancer Classification Using Transfer Learning

S. Dhage, H. Chawan, A. Hoskote, A. Dabholkar, V. Deshmukh

Proposed a DenseNet201-based CNN model achieving 90.39% accuracy in classifying 8 distinct types of skin lesions.

Education

Sardar Patel Institute of Technology

Bachelor of Technology in Computer Engineering

AlgorithmsSystem DesignOperating SystemsDistributed SystemsDBMSComputer Networks
9.01 / 10
Cumulative GPA
2021 — 2025

Get In Touch

Have an opportunity or just want to chat? Drop a line and I'll get back to you within 1–2 business days.