◇ process: whoami

software developer bridging engineering depth and business impact — cloud, AI, and products people actually use.

$ cat bio.md

I translate complex technical problems into scalable cloud and AI solutions that drive real adoption. I've shipped production features at Auvik Networks, built GCP backends at Auticon, and I care about making technology work for the people using it.

Toronto, ON · jaivshah.site

◇ jobs.d

jaiv@portfolio:~$ less ./experience.log

Production work across AI tooling, GraphQL microservices, GCP full-stack delivery, and test automation at scale.

[1] Software Developer — Auvik Networks

Sept 2025 – Present · Remote

Built an AI-powered Slack assistant using LangGraph + OpenAI API with a RAG pipeline (Pinecone vector DB) — eliminated hours of manual Slack searching for engineering teams
Reduced API response time 40% by migrating REST → GraphQL (Apollo Server, schema stitching across 3+ microservices)
Cut read latency 35% using CQRS + gRPC + Go in a multi-tenant microservice
Shipped reusable React/TypeScript component library (Hooks, Context, Storybook) used across product teams
Automated Helm + Kubernetes deployment pipelines, streamlining CI/CD workflows

[2] Full-Stack Developer — Auticon Canada

April 2025 – Oct 2025 · Remote

Owned full delivery of a location assessment service on Google Cloud Platform (GCP): Firebase Auth, Cloud Firestore, Cloud Functions — 99.9% uptime
Reduced data fetch latency 40% via RxJS observables + optimized Firestore queries
Achieved 60fps rendering with 1000+ concurrent map markers on mobile (viewport culling + requestAnimationFrame)
Technical bridge between engineering and business stakeholders; onboarded and mentored junior developer
Infrastructure provisioned with Terraform + AWS

[3] Backend Developer — Google Developer Student Clubs, McMaster

Sept 2024 – Apr 2025 · Hamilton, ON

Designed and shipped scalable secure APIs integrated with front-end products
Deployed Node.js microservices with Docker in Kubernetes environments

[4] Software Test Specialist Co-op — General Motors

Sept 2023 – Sept 2024 · Toronto, ON

Built company-wide test automation framework → accelerated releases 25%, cut manual testing 65%
Improved software reliability 40% through end-to-end UI automation with CI/CD integration
End-to-end Android testing using Robot Framework (Python) + Appium

guest@kernel:~$ ls ./projects --all

Internal AI tooling, a prior portfolio site, a PostgreSQL-backed CLI task tracker, and a vanilla JS drum pad — shipped with clear UX and practical automation.

production_ai

AI Slack Assistant

LangGraph + RAG over Slack history in Pinecone; cuts repeated questions and speeds onboarding for engineering teams.

LangGraphOpenAIPinecone

Ingestion: Historical Slack threads are chunked and turned into embeddings, then stored in Pinecone so semantically similar messages sit near each other in vector space.

Retrieval (RAG): When someone asks a question in Slack, the bot embeds the query, runs a vector search against Pinecone, and pulls back the most relevant past messages or docs as context — not the whole channel history.

Orchestration: LangGraph models the flow as steps (e.g. retrieve → optionally refine → call the OpenAI API with that context). That keeps behavior debuggable and lets you gate tool calls or retries.

Answer: The model generates a reply grounded in retrieved Slack content, so answers line up with how your team actually discussed the topic — reducing blind searching and repeated “has anyone seen…” questions.

legacy_site

Old portfolio

Earlier multi-page portfolio and project showcase hosted on Netlify — work samples, layout experiments, and links to builds like the drum pad demo.

HTMLCSSNetlify

↗ Live

cli_tool

TaskTracker

Lightweight CLI task manager for small teams: interactive menu or individual scripts for create, list, status, reassign, team reports, and deletes — backed by PostgreSQL with optional Slack webhooks for notifications.

PythonPostgreSQLSlack API

GitHub

Stack: Python 3.8+, PostgreSQL via psycopg, config with python-dotenv, HTTP to Slack with requests. A .env holds ConnectionURI and optional SLACK_WEBHOOK_URL (never committed).

Bootstrap: Run db_connection.py once to initialize the database schema before first use.

Interactive mode: tasktracker.py launches a text menu for creating tasks, listing, updating status, reassigning, viewing “my” tasks, team status, and bulk/demo helpers — all operations hit Postgres.

Script mode: The scripts/ folder exposes one file per action (e.g. create_task.py, list_tasks.py, update_status.py) so you can automate or wire into other tools without the menu.

Slack: slack_notifier.py posts to the incoming webhook when tasks are created, statuses change, or reassignments happen; if the webhook is unset, Slack steps are skipped cleanly.

web_app

JavaScript Music App

Tap colourful pads to trigger sounds and watch bouncing balls react to the beat — vanilla JS, HTML5, and CSS3 only.

JavaScriptHTML5CSS3

↗ Live GitHub

ls -R /tech_stack

📁 frontend/

01 react_18.tsx

02 next_app.router

03 typescript_strict.ts

04 storybook.chromatic

05 rxjs_streams.pipe

📁 backend/

01 node_runtime.bin

02 apollo_graphql.gateway

03 grpc_services.proto

04 go_microsvc.main

05 kafka_topics.stream

📁 infrastructure/

01 gcp_firebase.tf

02 kubernetes.helm

03 terraform_modules.hcl

04 docker_compose.yml

05 github_actions.ci

◇ education.sys

jaiv@portfolio:~$ cat degrees.txt

Bachelor of Technology in Software Engineering — McMaster University (2023–2025) · GPA 3.8/4.0

Advanced Diploma in Computer Engineering — Humber College (2018–2021) · Dean's Honors List

sh contact.sh --run

establish_connection

Hiring, collaboration, or questions about my work — use the protocols or send a payload through the form.

protocol: smtp → protocol: github → protocol: linkedin →