Internal documentation pipelines
Unstructured docs, tickets, and tribal knowledge piped through n8n + AI extraction into a clean, queryable source of truth.
- n8n
- OpenAI / Claude
- Vector DBs
Ahmedabad, India · available for new engagements
I build n8n workflows that permanently eliminate repetitive SaaS ops work.
Not templates. Not tutorials. Scoped, built, and handed off — so your team stops doing the same thing twice.
- Years shipping production systems
- 5+
- Edge environments deployed
- 10+
- n8n workflows shipped
- 50+
- Languages spoken fluently
- 3
The problem
The ops work that should take 20 minutes takes 3 hours. Every week. Because it's manual, undocumented, and nobody has time to fix it.
About
Operator first, automator by habit.
I started by writing firmware for sensors that sit on rooftops in 40°C heat and report air quality every 30 seconds. When your code runs on 1,000 boxes you can't physically reach, you learn fast that anything manual will eventually break at 2 AM.
That instinct carried into product ownership. Every time I saw someone on my team doing the same copy-paste ritual twice a week, I'd stop the meeting and go build a workflow for it. Over three years that turned into an entire ops layer — firmware rollouts, documentation pipelines, KPI reporting — that quietly removed a dozen recurring chores from our calendar.
Now I do that work directly for other teams. I don't write thought-leadership, I don't sell courses, and I don't build templates. I ship the specific workflow your team keeps putting off, hand it over fully documented, and get out of the way.
— Joy Jacob, Ahmedabad
What I build
Three things, done properly.
Edge deployment workflows
Proven across 10+ environments. Firmware rollouts, device provisioning, and config drift checks that run without a human in the loop.
- n8n
- MQTT
- Python
- Shell
BI pipelines from unstructured data
Emails, PDFs, CSVs, API dumps — reshaped into the dashboards and alerts your ops team actually reads.
- n8n
- Google Sheets
- Postgres
- AI extraction
Selected work
Systems shipped in the last 12 months.
Descriptions generalized to respect client confidentiality. Happy to walk through specifics on a call.
Environmental SaaS
Edge deployment across 10+ environments
Problem
Firmware and config rollouts to field devices were manual, error-prone, and blocked product releases.
Approach
n8n-orchestrated deployment pipeline tying together device registry, OTA triggers, and health checks — with a rollback path.
Outcome
Release cycle shortened, deployment errors near-zero, ops team freed from weekly firefighting.
- n8n
- MQTT
- Python
- Edge hardware
Internal ops
AI-assisted documentation pipeline
Problem
Product knowledge scattered across Slack, email, and engineering notes. New hires took weeks to ramp.
Approach
n8n workflow ingests unstructured sources, uses AI extraction to normalize, and pushes into a searchable internal doc layer.
Outcome
Documentation stays fresh without a dedicated writer. Ramp time measurably shorter.
- n8n
- LLM extraction
- Notion / wiki backend
Business intelligence
BI from unstructured inputs
Problem
Leadership wanted weekly KPIs that lived across emails, PDFs, and third-party dashboards nobody owned.
Approach
Scheduled n8n pipelines pull, parse, and push into a single warehouse + one-page dashboard.
Outcome
Monday-morning KPI readout runs itself. No more manual spreadsheet stitching.
- n8n
- Google Workspace
- Postgres
- Metabase
How I work
Built to run without me.
-
01
Scope
One call to find the real bottleneck. Concrete deliverable, not a retainer.
-
02
Build
n8n workflow + whatever glue it needs. Reviewed with you at milestones.
-
03
Hand off
Your team owns it. Credentials, runbooks, failure modes documented.
-
04
Support
Optional. Most clients don't need it — the workflow just runs.
Principles
Four things I won't compromise on.
Scoped, not subscribed
Flat-rate projects with a defined endpoint. No open-ended retainers unless the workflow genuinely needs ongoing maintenance.
Documented by default
Every handoff includes node-by-node runbooks, credential inventory, and failure playbooks. Your team owns it the day it ships.
Built on your stack
Workflows run on your n8n instance, your cloud, your credentials. No vendor lock-in to me.
Honest about fit
If automation is the wrong answer, I'll tell you on the scoping call. No half-fit engagements.
Background
Five years of shipping production systems.
From firmware on physical devices in the field to product ownership of an environmental monitoring SaaS — the pattern is the same: find the friction, remove it, move on.
- Jan 2024 — Present
Product Owner · Oizom
Environmental monitoring SaaS. Roadmap ownership, cross-functional delivery across firmware, cloud, and field ops. Ship cycle cut from quarterly to monthly releases.
- Jun 2021 — Dec 2023
Embedded Engineer · Oizom
Firmware in C/C++ on ESP32 and STM32. Python-based QC automation for production lines, DWIN DGUS hardware UIs, PCB review with the electronics team.
- Oct 2020 — May 2021
Technical Project Manager · Jayso Labs
Ran AI / Python education project delivery. Hands-on engineering review, client scoping, and curriculum design.
- Jan — Apr 2020
Firmware Intern · Oizom
First exposure to production firmware. LoRa, WiFi, and BLE stacks on resource-constrained controllers.
- 2019
PCB Design Intern · Odinub
Schematic + PCB layout for IoT dev boards. KiCAD / Eagle. Shipped hardware that got into customer hands.
- 2018
IoT Intern · techiesms
Built end-to-end IoT projects for a YouTube electronics channel with 1M+ subscribers. First time writing code real people would run.
Education
- BE, Electrical & Electronics Engineering · L.D. College of Engineering (2017–2020)
- Diploma, Electronics & Communication · Government Polytechnic Gandhinagar (2014–2017)
Recognition
- Smart India Hackathon 2018 — Grand Finalist
- Best Academic Student Award
Stack
Tools I reach for.
Primary
- n8n
- Generative AI tooling
- Server / infra ops
Languages
- Python
- C / C++
- App Script
- Shell
Protocols & Hardware
- LoRa
- MQTT
- UART
- I2C
- WiFi
- BLE
- REST / HTTP
Tools
- VS Code
- GitHub
- Node-Red
- KiCAD
- Eagle
- DWIN DGUS
Spoken
- English
- Hindi
- Gujarati
Also building
Dhumketu.
A venture focused on the overlap of AI, automation, and ops infrastructure. Early-stage — more to share soon.
FAQ
Things people ask before booking.
What does a typical engagement look like?
A 20-minute scoping call, a fixed-scope proposal within 48 hours, build in 1–4 weeks depending on complexity, then a handoff session. No retainers unless you specifically want one.
Do you work on retainer?
Only if the workflow genuinely needs ongoing changes — e.g., it integrates with APIs that break, or you want a shared backlog. Most projects don't need one. I'd rather build something that runs itself and move on.
Who owns the workflow after handoff?
You do. Credentials, node configs, runbooks, and failure modes all live in your accounts. You can fire me tomorrow and everything keeps running.
Can you work with tools other than n8n?
Yes — Zapier, Make, Node-Red, or custom Python if the problem fits better elsewhere. n8n is the default because it's self-hostable and the node graph is easier for ops teams to read than code.
Do you do ongoing support?
Optional hourly bucket post-handoff. Most clients don't pull from it. The workflow is documented well enough that their own team handles changes.
What's out of scope?
Full-stack product builds, mobile apps, design work, or anything customer-facing. I ship internal ops infrastructure — the plumbing, not the storefront.
Contact
What repetitive work is eating your week?
Tell me in two lines. If I can help, we'll set up a 20-minute call.