NEWHow a B2B SaaS went from 4 to 131 AI answers citing it

Co-founder and CTO · TechDevs

Prasad Iyer

Automated his own company first. Now he builds yours.

Former Oracle applications developer and author of more than 100 AI skills and 40+ MCP servers. Prasad builds the AI agents and automations TechDevs ships, with the discipline banks demand of their models: every output traceable, every release approved.

Find him on
Prasad Iyer, co-founder and CTO of TechDevs, AI and operationsEx-Oracle · AI systems builder

Prasad IyerCo-founder · CTO, AI & Operations

AI skills written
100+
MCP servers built for businesses
40+
applications developer at Oracle
2.5 yrs
co-founded TechDevs
2020

The origin

Two full-time jobs. Zero spare hours.

prasad@techdevs: ~Profile

prasad ›

Profile: whoami --career. College: writing code, then freelancing for paying clients. Oracle: applications developer, about two and a half years. Machine learning projects for model governance at top banks. Ran TechDevs the whole time: two full-time jobs.

Profile: ls ~/skills | wc -l. 100+ skills, written and in daily use. SEO audits, site builds, proposals, outreach, video. Run in Claude, ChatGPT, Gemini and other agents.

Profile: agents --status. 40+ MCP servers: business operations, run by any AI agent. Agency OS running TechDevs day to day. Anything that sends or spends waits for a person.

Prasad started writing code in college and was doing client work before he graduated. That practice became TechDevs.

After college he joined Oracle as an applications developer and stayed for about two and a half years. He never stopped running TechDevs. For that whole stretch he held two full-time jobs.

There was only one way to make that work. Anything that could become a system became a system: delivery, reporting, follow-ups, sales. He automated his own company because he had no other option, and that constraint is exactly what he builds for clients now.

“It was the only way the arithmetic worked.”
01 / 05

Coder, then freelancer

Wrote code for paying clients while still studying. Websites, apps and internal systems, shipped often enough to learn what works and what only demos well.

  • Full-stack
  • Client work
02 / 05

Applications developer

Worked on machine learning projects used by top banks for model governance: the controls that decide which models a bank may run, and prove why. Enterprise code, enterprise review, enterprise consequences.

  • Machine learning
  • Model governance
  • Banking
03 / 05

Two full-time jobs

Ran TechDevs alongside Oracle. Automated delivery, reporting and sales so the company kept moving without him at a desk.

  • Automation
  • Operations
04 / 05

Co-founder and CTO

Co-founded TechDevs in 2020 with Prithvi Jadwani. 130+ clients across India, the Gulf, the US, Canada and Australia.

  • Engineering
  • AI strategy
05 / 05

The AI engine

More than 100 AI skills, 40+ MCP servers that let businesses run their operations from Claude, ChatGPT, Gemini or any AI agent, the agency operating system TechDevs runs on, and production agents for clients.

  • AI agents
  • Skills
  • MCP

Where the discipline comes from

Banks cannot run a model nobody can explain.

At Oracle, Prasad worked on machine learning projects that top banks used for model governance. In a bank a model does not go live because it looks right: it is inventoried, validated, approved and watched. He builds agents the same way. Pick a control to see what it became.

  • Agents run on named, versioned skills, so you always know which playbook produced an output and can change it in one place.

    Browse the skills
  • Skills are benchmarked: the same task, the same model, a plain prompt against the skill, graded blind. The numbers sit on each product page with the method.

    See a benchmark
  • Numbers an agent drafts point back to the document, page and line they came from. A claim without a source does not leave the system.

    See the evidence trail
  • Anything that sends, spends or files waits in a review queue. Agents prepare; a person approves.

    See the agents
  • Before go-live, agents run side by side with your team on last month’s real inputs, and accuracy is measured rather than assumed.

    How a build runs

AI, every day

A power user who writes the playbooks.

Prasad works with AI agents all day: Claude, ChatGPT, Gemini and whichever model fits the step. He does not prompt his way through a job; he writes the skill once, tests it on real work, and lets any agent run it the same way every time.

100+
AI skills in his library
40+
MCP servers built so businesses run operations from Claude or any AI agent
8
agent types shipped, one review-gate model
01

Skills

100+ written playbooks: SEO audits, site builds, proposals, video, outreach, code review. They run in Claude, ChatGPT, Gemini and other agents.

02

Parallel subagents

Specialist agents fan out across isolated worktrees; a reviewer checks the merge.

03

40+ MCP servers

Your systems exposed as tools any AI agent can call, from Claude to ChatGPT to Gemini, scoped, with no send button unless you add one.

04

Agent SDK workers

Long-running jobs on a queue, each with its own reviewer.

05

Hooks and hard rules

What an agent may never do, enforced in code before it tries.

06

Review loops

Every change loops through automated review until it comes back clean.

automate-seoseo-audit-reportbest-webdevbusiness-site-builderproposal-deckonboard-clientreviewloopblog-image-gengmb-optimizercrawl4ainano-bananasmooth-landing-pageseo-rank-analystseo-link-intelseo-backlink-prospect-finderseo-content-scorerseo-duplicate-analystseo-report-narrator
seo-page-rebuilderseo-outreach-prospectorseo-ai-visibilityseo-blog-writerseo-crawl-workerwebdev-strategistwebdev-copywriterwebdev-section-builderwebdev-svg-artistwebdev-video-producerwebdev-image-directorwebdev-site-analystwebdev-qanew-featurecode-structureevidence-driven-testingbefore-and-afterseo-technical

A slice of the library. Some of these are for sale. Browse the store

Systems he shipped

In production, not in a deck.

Client names are hidden on purpose. Stages are stated honestly.

LiveD2C gourmet food brand

An AI concierge that sells from the catalogue

A commerce agent that answers, recommends and sells around the clock, running on the client’s own infrastructure.

  • WhatsApp Business API
  • Shopify Storefront API
  • Vector knowledge base
  • AWS
Active buildMulti-channel ordering platform

Agents that write orders back into a live store

Agents writing into a live commerce system, not just reading from it.

  • Shopify Admin API
  • Order rules engine
  • Dashboard
Live, our productOur own product

Agentic Outreach: signal-driven LinkedIn and email

We run agents on ourselves first. The grounding checks exist because we caught the failure.

  • Claude
  • MCP
  • LinkedIn
  • Email
Live, used dailyOur own agency

TechDevs Agency OS

The same method we sell, used every day on our own revenue.

  • Claude Agent SDK
  • Job queue
  • Postgres
  • Review gates
Shipped, in useFounder-led brand

One prompt to a finished reel

An AI video pipeline with a human gate, built and used by the person selling it.

  • Claude
  • Voice clone
  • Avatar
  • Auto-edit

Every case, with what each one proves

How he works

Four rules he does not bend.

  1. 1

    Automate yourself first

    He automated his own company before anyone else’s. Most of what TechDevs sells started as something he needed.

  2. 2

    Evidence or it does not ship

    If a number cannot be traced to its source, the agent does not get to say it.

  3. 3

    Agents prepare, people approve

    Anything that sends, spends or files waits for a person. Speed never skips the gate.

  4. 4

    You get the keys

    He ships the infrastructure, documents it, and hands it over. You own it.

“Most companies think about automation as a way to spend less. The interesting version is leverage: the same team holding more ground, without hiring for any of it.”
Prasad Iyer

Client love

Clients on what the team has built.

4.9Average client rating

They automated my registrations and even did SEO beyond the proposal. A fantastic job.

Rima RohraYoga teacher & sound healer · Yog with Rima

“A colleague recommended Prithvi for our landing page design, and it was the best decision we made. Within a month, we saw a 40% increase in conversions.”

Sujal MehtaLocal Guide · Google review

“I’ve worked with Prithvi and his team at Tech Devs multiple times, and every single experience has been exceptional… once it starts showing results, the visibility and impact are incredible.”

Samelip ShahLocal Guide · Google review

“One of the most helpful guys I’ve met when it comes to seo. They’re well equipped, and have proven strategies for fintech.”

Siddharth ThakkarGoogle review

“Got my website done from tech devs. They were swift and finished it before time. Very humble courteous team with great knowledge for problem solving.”

Manoj BodhwaniGoogle review

“Love the energy and honest recommendations. Also the work got completed on time. Thank you thank you thank you!!!!”

Rima RohraGoogle review

“Great SEO services! Highly Recommend!!”

Devika MandyaniGoogle review

FAQ

Questions people ask first.

What did Prasad do at Oracle?

He was an applications developer for about two and a half years, working on machine learning projects that top banks used for model governance. He ran TechDevs at the same time.

Will Prasad build it himself, or hand it to a junior?

He leads the technical strategy on every AI engagement and ships the infrastructure himself. Once it is live, the team he hired and trained maintains and extends it.

What does "100+ AI skills" actually mean?

A skill is a tested playbook an AI agent follows: instructions, scripts and checks for one job, such as an SEO audit or a proposal deck. They work with Claude, ChatGPT, Gemini and other agents. Prasad has written more than a hundred, and TechDevs delivers client work with them every day.

What is an MCP server, and why build 40 of them?

MCP (Model Context Protocol) is the standard way to give an AI agent tools: read the CRM, draft the invoice, check stock, update the sheet. Prasad has built more than 40, so a business can run its operations from Claude, ChatGPT, Gemini or any agent that speaks MCP, with nothing that sends or spends going out without approval.

What happens on the first call?

Thirty minutes on one workflow. He finds what is actually worth automating, says so honestly when nothing is yet, and you leave with a build plan either way.

Do we own the system, or rent it from you?

You own it. He ships the infrastructure and hands you the keys, with his team maintaining it for as long as you want them to.

My problem is demand, not systems. Who should I talk to?

Prithvi Jadwani, his co-founder, who runs the organic engine: SEO, AI-search visibility, Reddit and community growth, local SEO, and CRO.

Follow the work

Where Prasad publishes.

Prasad builds in public: AI agents, skills, MCP servers and the operations behind them, shared as the systems ship.

Meet his co-founder, Prithvi Jadwani, who runs the organic engine. Or read how the two of them built TechDevs.

What do you want to automate?
Budget for a first build
Timeline
Best way to reach you

Prasad reads every request himself. No sequences, no newsletter.