Skip to content

Available for full-time roles from January 2027

Boston, MA--:--:--

full-stack software engineer

I build platforms, APIs, pipelines, dashboards, infrastructure
that hold at scale.

Email SmitRésumé.pdf

Experience

Shipped to production.

Three roles, three production codebases. Each one replaced something manual with something that runs on its own.

01 / 03

→

Maryland, USA

Current

Origin Wireless

Software Engineer Co-op

Internal tooling and fleet observability for a wireless-sensing company. Shipped operator-facing dashboards, access control, and the monitoring layer under them.

  • Next.js
  • Redis
  • BullMQ
  • MySQL
  • Apache Casbin
  • Jest

  • Built a dashboard that replaced a manual curl-based terminal workflow for constructing and sending network topologies, cutting task completion time and eliminating a recurring source of operator error.

    2 min30 sectask completion

  • Designed and shipped role-based access control (admin, partner, maintainer) using Apache Casbin, closing a security gap by restricting partner visibility into sensitive model configuration without breaking internal team workflows.

  • Independently built a fleet monitoring dashboard in Next.js, using Redis, BullMQ, and MySQL to aggregate CSI status across devices via background jobs, triggering real-time alerts on service failures and supporting a company-wide pilot rollout.

  • Authored setup and usage documentation for the fleet dashboard codebase to support onboarding and cross-team adoption.

  • Revamped an internal service dashboard with updated branding and optimized rendering performance, and built new dedicated dashboards for multiple services, supporting cross-team monitoring.

  • Wrote unit and integration tests with Jest covering core frontend and backend workflows end to end, catching regressions before code reached production.

02 / 03

→

Ahmedabad, India

Jayso Labs

Software Engineer — Founding Engineer

First engineer on a consumer platform for exploring city development plans and real estate projects across India. Owned architecture, infrastructure, and the geospatial layer.

  • Node.js
  • AWS S3
  • CloudFront
  • Docker
  • AWS EKS
  • Terraform
  • BigQuery GIS

  • As founding engineer, architected and built the core platform from the ground up, scaling it to over 1 million active users exploring city development plans and real estate projects across India.

  • Built a distributed map tile server using Node.js, AWS S3, and CloudFront, designed to support additional and future city expansion maps layered onto Google Maps, using caching and CDN delivery to keep tile loading fast and efficient at scale.

  • Containerized the platform with Docker and deployed it to AWS EKS, writing Terraform scripts to provision infrastructure as code, enabling repeatable, scalable deployments for the map tile server and supporting services.

  • Designed a geospatial query API on top of BigQuery's GIS functions that returns a user's nearest projects and land listings first, with cursor-based pagination and multi-layer caching to keep results fast and consistent as usage scaled to millions of requests.

03 / 03

→

Ahmedabad, India

Oizom Instruments

Software Engineering Intern

Environmental sensing hardware company. Automated the infrastructure under internal tooling and built the calibration dashboard the hardware team used.

  • Terraform
  • Packer
  • AWS EC2
  • Next.js

  • Automated EC2 infrastructure provisioning using declarative Terraform configurations and custom Packer AMIs, eliminating manual setup for internal tool deployment.

  • Built a Next.js sensor calibration dashboard that replaced manual validation checks with automated calibration tracking, reducing validation overhead.

    baseline−30%validation overhead

Selected projects

Built, then deployed.

Built outside of work, and deployed rather than demoed. Every one of these has a repository or a running URL behind it.

01

5live apps

Agent Arena

Two AI models race side by side on the same real multi-app task, and the server, not the model, decides who actually succeeded.

  • Built a split-screen race where two Gemini agents act on real GitHub, Notion, Slack, Gmail and Google Calendar accounts through Composio, with every tool call streamed live over Server-Sent Events and replayable from Postgres.
  • Wrote per-task verifiers that check the real outcome directly against each app with no model involved, so a model claiming "done" counts for nothing; the first race caught the stronger model reporting success with the wrong issue on its page.
  • Made it hold up on serverless free tiers: SSE tails the events table by id and resumes from Last-Event-ID, 429s retry with backoff and jitter, and compacting tool results cut a GitHub payload by about 80% before it reaches the model.
  • Next.js
  • TypeScript
  • Composio
  • Gemini
  • Supabase
  • SSE
  • Vitest
Live demoRepository
02

4apps, read-only

The Morning Edition

An AI editor that reads your Gmail, Calendar, GitHub and Slack each morning and sets what matters as a broadsheet front page.

  • Designed a three-stage gather, edit and publish pipeline: four apps fetched in parallel with per-app timeouts, Gemini returning Zod-validated JSON with one corrective retry, and a grounding pass that drops any story citing a source that doesn't exist.
  • Enforced a read-only guarantee in code three ways: a static tool allowlist on a single execution path, a runtime check that fails closed unless Composio marks the tool read-only, and read-only OAuth scopes for Gmail and Calendar.
  • Added multi-user sign-in with Supabase magic links, each reader connecting their own accounts through Composio's hosted auth, with every edition stored as JSON so it reprints exactly as it ran.
  • Next.js
  • TypeScript
  • Composio
  • Gemini
  • Zod
  • Supabase
Repository
03

Google Workspace MCP Server

An MCP server that lets Claude act as an autonomous agent across Gmail, Drive, and Calendar.

  • Built an MCP server enabling Claude to act as an autonomous agent across Gmail, Drive, and Calendar, exposing OAuth2-secured tools so LLMs can execute real workspace actions rather than just generate text.
  • Node.js
  • Google APIs
  • OAuth2
  • Claude AI
  • MCP
Repository
04

5Kreq/sec

Cloud-Based API Deployment & Infrastructure Automation

A distributed AWS backend built for load, provisioned entirely as code.

  • Architected a scalable, resilient distributed backend handling 5K requests/sec on AWS using EC2, ALB, RDS, and S3, with zero-downtime deployments demonstrating cloud-native design at production scale.
  • Automated cost-effective infrastructure provisioning using Terraform, Packer, and GitHub Actions, cutting manual deployment time to near-zero and enforcing infrastructure-as-code across all environments.
  • Node.js
  • AWS
  • Terraform
  • Packer
  • GitHub Actions
Repository
05

10Kevents/sec

Logflow Pipeline

A cloud-native distributed streaming system for log analytics, fault-tolerant by design.

  • Designed and built a fault-tolerant distributed streaming system using Kafka and Spark on GCP, processing 10K log events/sec with sub-second latency, persisting analytics to PostgreSQL, and automating infrastructure with Terraform.
  • Kafka
  • Spark
  • PostgreSQL
  • Terraform
  • GCP
Repository
06

Codial

A real-time code submission platform with sandboxed multi-language execution and automated grading.

  • Developed a real-time coding platform using Next.js, DynamoDB and CodeMirror, deploying the frontend on Vercel and running Judge0 in Docker containers on AWS EC2 for scalable multi-language code execution and automated grading.
  • Next.js
  • DynamoDB
  • Judge0
  • TypeScript
  • Docker
Live demo

Stack

Front to back, end to end.

Everything Smit builds with, grouped the way he reaches for it — from the React on the screen to the Terraform under the cluster.

Technical skills, grouped by area
AreaTools
01Languages
  • JavaScript
  • TypeScript
  • Python
  • Java
  • SQL
  • Go
02Frontend
  • React
  • Next.js
  • Redux
  • Tailwind CSS
  • Material UI
  • Chakra UI
03Backend
  • Node.js
  • Express
  • REST APIs
  • GraphQL
  • FastAPI
  • Django
  • Flask
  • Microservices
  • WebSockets
04Cloud & Infrastructure
  • AWS EC2
  • S3
  • RDS
  • Lambda
  • ALB
  • CloudWatch
  • GCP
  • Docker
  • Kubernetes
  • Terraform
  • GitHub Actions
  • Jenkins
05Data & Messaging
  • PostgreSQL
  • MongoDB
  • MySQL
  • Redis
  • Cassandra
  • Kafka
  • Spark
06AI / ML
  • LangChain
  • ChromaDB
  • RAG Pipelines
  • Agentic Workflows
  • GenAI
  • LLMs
  • GPT APIs
  • PyTorch

Education

Northeastern University
Master of Science in Software Engineering Systems
3.77GPABoston, MA
Nirma University
Bachelor of Technology in Electronics and Communication Engineering
3.7GPAAhmedabad, India

FAQ

Straight answers.

The questions that actually get asked, answered without hedging.

Q01Who is Smit Patel?
Smit Patel is a software engineer based in Boston, MA, completing an MS in Software Engineering Systems at Northeastern University in December 2026. He was the founding engineer at Jayso Labs, where he architected a platform that reached over 1 million active users, and is currently a software engineering co-op at Origin Wireless.
Q02What kind of engineering work does he do?
Full-stack work weighted toward infrastructure and distributed systems: Node.js and Next.js services, AWS and GCP, Kubernetes and Terraform, Kafka and Spark streaming, and geospatial APIs on BigQuery GIS. He has repeatedly owned systems end to end rather than contributing to one layer.
Q03What is he looking for?
Full-time software engineering roles in the United States starting January 2027, after he graduates from Northeastern in December 2026. He is open to backend, infrastructure, platform, and full-stack positions.
Q04What is the largest system he has built?
The Jayso Labs platform, which he architected and built from the ground up as founding engineer and which scaled to over 1 million active users. It included a distributed map tile server on Node.js, AWS S3, and CloudFront, and a geospatial query API on BigQuery GIS handling millions of requests with cursor-based pagination and multi-layer caching.
Q05What languages and technologies does he use?
JavaScript, TypeScript, Python, Java, SQL, and Go. React and Next.js on the frontend; Node.js, Express, FastAPI, Django, and GraphQL on the backend; AWS, GCP, Docker, Kubernetes, and Terraform for infrastructure; PostgreSQL, MySQL, MongoDB, Redis, Cassandra, Kafka, and Spark for data; and LangChain, RAG pipelines, and agentic workflows for AI work.
Q06How do I contact him?
Email patel.smit11@northeastern.edu or call 857-397-9736. He is also at linkedin.com/in/smitpatel3107 and github.com/SmitPatel-31.

Longer write-ups on the systems above live in the writing section.

Available for full-time roles from January 2027

Let's build
something.

Email — the fastest way inpatel.smit11@northeastern.edu