Engineer · Founder · Agentic AI systems

I build AI that does the actual work.

I take the process quietly bleeding a business time or money, design the agentic system around it, and ship it live. Voice agents that answer the phone and book the appointment. Software that reads a contract and writes the offer. A brain that runs an entire sales floor from Slack. A team of coding agents that lands pull requests on a client's production platform. Twenty-plus systems shipped across five countries, many still running in production.

Book a 30-minute call See the work

20+Systems shipped

5Countries shipped to

114Pull requests landed in eleven weeks

356/356Tests green on a critical audit

5x faster fulfillment 2-3 hrs → 5 min CRM entry 1,578 → 410 events scored 7,842 accounts qualified 25-35 min → 5 min per offer 1 hr → 10 min per proposal 26 missing contracts found 93-page plan set read

Where I come from

Frontier AI lab

I work for a top frontier AI lab

I grade its coding agents, build training data, and build the environments its models train in. I can't name the lab here. The work keeps what I ship current with where the models actually are.

Payments engineering

Interned at PhonePe

I interned at PhonePe, India's largest payments platform, working on lending analytics on production data. I turned down the return offer to build on my own. Production data that moves real money set the bar I hold my own work to.

Owner-operator

Founder of an AI studio

I run a small AI automation studio. Four of us have delivered for more than 20 businesses, and I ship as the builder, hands on the keyboard. I own the outcome, sit in production, and fix it until it holds under real load.

Selected work

Across construction, healthcare, finance, sports, and services. Client names are held back under confidentiality. The systems and the numbers are real.

Agentic products & brains

AGT · 01In production01

Builder / architect · In production

A software factory of coding agents

A harness that runs a team of coding agents, each with different permissions. A scout with no edit tools, a builder that never pushes, a drafter with no shell, and a verifier whose only job is to find fault with the others. A hook makes every session tag each claim as checked live, remembered, or a guess. It landed 114 pull requests on a client's production platform in eleven weeks.

114 PRs in 11 weeksClaude CodeHooksMulti-agent

AGT · 02Live02

Nashville venture studio · In production

A business-development brain in Slack

Drop a business card, a voice memo, or a conference recap into Slack, or forward an email, and it extracts the contacts, enriches the leads, writes the CRM record and a per-contact dossier, and drafts the follow-up in the founder's own voice. Three pipelines on n8n, roughly forty-seven nodes.

2-3 hrs → ~5 minn8nClaudeDeepgramAttio

AGT · 03In production03

Finance & real estate · In production

A company brain for finance teams

It reads a company's own ERP, accounting file, bank statements, and spreadsheets read-only, returns one trustworthy set of numbers with every figure traceable to its source, and catches the wrong one before it gets filed.

RAGDocument parsingReconciliation

AGT · 04Shipped04

Multi-agent publishing · Shipped

A scout, writer, and editor agent chain

A scout agent researches, a writer drafts, and an editor reviews. The chain researched, wrote, and published a post to a live site on a schedule, with nobody in the loop.

Nobody in the loopAgent chainsScheduled runs

Voice & conversational agents

VOX · 05In production05

Spinal surgeon, UAE · In production

Consultation recordings to finished clinical reports

It turns a doctor's consultation recording into a finished clinical report. Built for a spinal surgeon in the UAE, who still uses it. Voice booking, patient scheduling, and an AI chat agent sit around it.

VAPIRetellLiveKitDocument AI

VOX · 06Live06

Dental clinic, Puerto Rico · Live

A bilingual WhatsApp concierge

Spanish-first over WhatsApp. It books, reschedules, and cancels straight into the calendar, handles the local insurers, transcribes patient voice notes, redacts protected health information, and routes emergencies after hours, on a real clinic number.

Twilio WhatsAppClaudeSupabaseCalendar

VOX · 07In production07

US pharmacy · In production

An outbound voice agent for a pharmacy

HIPAA-compliant. It bridges the pharmacy's software to a voice agent that pulls the patient, prescription, and insurance data, places the call, verifies identity, delivers the message, and captures the outcome across seven paths and twenty-plus edge cases.

RetellElevenLabsNodePharmacy API

VOX · 08Live08

Sports-facility platform · In production

An embeddable voice-and-chat booking agent

Website visitors book racquet-sport courts over both text and live voice from a single script tag. A backend proxy protects the key and stays the single source of truth for config, so it drops onto any site cleanly.

VAPIReactExpressRailway

VOX · 09In production09

Wellness clinic · In production

A full AI reception system

An inbound and outbound voice agent for a wellness clinic, wired to a GoHighLevel CRM with quiz-based segmentation, intake forms, payment sync, and compliance handling. It replaced manual phone reception end to end.

VAPIRetellGoHighLevel

Products & platforms

PRD · 10In production10

Inbox infrastructure company · In production

Fulfillment automation + a custom SaaS dashboard

I automated an inbox-setup company's entire fulfillment flow and built them a SaaS dashboard for full operational visibility. Fulfillment time dropped fivefold and they added five figures in monthly recurring revenue within weeks, delivered in under four weeks.

"They streamlined our entire process and got it live in under 4 weeks. We've already added five figures in MRR thanks to these guys."Co-founder, inbox infrastructure client

5x faster fulfillmentSaaS dashboardn8n

PRD · 11In production11

Lending / fintech · In production

A self-serve voice-demo product for lenders

A prospect requests a demo call, standard or custom, and can upload their own FAQ, product, and flow documents to spin up a personalized voice demo on the spot. I built the product and the demo funnel behind it.

Next.jsVoice AITypeScript

Revenue & business systems

OPS · 12Delivered12

80-person US expense management firm · Delivered

A contract register inside the CRM

The firm tracked 168 renewals by hand. I built a contract register inside its CRM in 12 days, and it found 26 signed agreements missing. Every extracted term carries the clause it came from, and the build fails on any excerpt that is not verbatim. The work became a service, Fiduka.

26 missing agreements foundCRMDocument extractionClause citations

OPS · 13In production13

Transport & traffic engineering firm · In production

A systems-integration audit and RevOps build

I mapped the current-state workflows across every service line of an engineering firm, then automated their fee-proposal process across seven templates and designed HubSpot dashboards giving leadership real-time visibility into pipeline, revenue, and utilization.

1 hr → 10 min per proposalHubSpotProcess mapping

OPS · 14In production14

Landscaping firm, Atlanta · In production

A revenue-recognition rebuild on HubSpot

For a company running about 2,600 live deals and recurring-payment automations. I built a revenue-recognition engine and the deal-type and lost-reason model as additive, non-destructive changes, recognizing revenue in the month the service is delivered.

2,600 live dealsHubSpotRevOps

GTM & data engines

GTM · 15Live15

Venture studio · Live

An event-discovery engine across 25 metros

Four scrapers pull founder and startup events across twenty-five US metros. The pipeline dedupes, scores each one zero to a hundred with an LLM, enriches to clean structured data, and mails out what clears the bar. One run took 1,578 in and shipped 410. Backed by 57 tests.

1,578 → 410 scoredPythonPostgresGemini

GTM · 16In production16

Multi-client · In production

A lead qualification & signal engine

Point it at any B2B list. It grades every account against an ideal-customer profile, enriches the survivors with public signals from their site, hiring pages, maps, and news, then scores and tiers them A through F. One run took 7,842 accounts to a clean set.

7,842 accounts scoredPythonICP scoring

GTM · 17Live17

Growth agency · In production

A signal-first cold-email engine

It scrapes buying signals before anything else, a new exec in role, headcount growth, recent funding, live hiring pages, then buckets accounts by pain, writes copy per bucket, and extracts emails last. Free-first by design and hard-capped at five dollars of scraping per three-to-four-thousand-lead run.

$5 cap / 3-4K leadsPythonn8nPerplexity

GTM · 18Running18

Content & outbound · Running

Content and cold-email engines

An always-on pipeline that watches eleven feeds, writes opinionated takes with matching imagery, and posts them for a human to approve before anything ships. A sixty-five-node workflow, plus deliverability-tuned outbound that gets read.

n8nGeminiInstantly

GTM · 19Built19

Scraperly · College project · Built

Scraperly, a contact data engine

I spent three months studying how a large contact database spots bots. Then I built an engine that works out each company's email pattern, and browsers that move like people, with random pauses, rotating sessions, and proxies. It pulled thousands of verified leads a day.

Thousands of verified leads a dayBrowser automationProxiesEmail patterns

Real estate & construction

REI · 20In production20

Denver house-flipping investor · In production

Offer automation for a real-estate investor

I studied his seven process walkthroughs and turned a twenty-five to thirty-five minute per-property offer process into a five-minute conversation. Commands run his offer calculator against his own workbook, log the deal, fill the Colorado purchase-agreement PDF field by field, and draft the agent email.

25-35 min → 5 minClaude CodePythonPDF autofill

REI · 21In production21

Real estate developers, North India · Live

An AI-native construction OS

A construction operating system set up around a developer's own workflows and live in 30 days. I have done this for two real estate developers in North India. More at NeevOS.

Live in 30 daysAgentic pipelinesWorkflow mappingFull-stack

REI · 22Shipped22

Construction-tech platform · Shipped

Three tools for a construction-tech platform

AI Takeoff reads a full construction plan set, tested on 93 pages, finds the schedule pages, and cites the spec requirement behind every match. A bill of materials tool builds the bill for a job. An insurance certificate tracker tracks every subcontractor's certificate and chases the ones about to expire.

93-page plan setDocument AISpec citationsAutomated follow-ups

Security & production safety

SEC · 23In production23

PE-backed US energy firm · In production

Security audit & hardening of a deal-pricing engine

A hundred-thousand-line trading app. A fifty-six-finding audit, every finding tied to a file and a line across six rounds of adversarial verification, then around twenty-nine fixes with the suite green at 356 of 356. One catch: a fee bug that would have stopped any deal over seventy million dollars from reaching a human.

356/356 tests greenTypeScriptThreat modeling

SEC · 24In production24

250-person consulting firm · In production

A production-safety layer for an internal AI app store

An internal store where a 250-person team browse, submit, and review their own AI tools under role-gated approval. I own the safety layer, a review harness and hard rules on every pull request, so non-engineers ship internal tools without breaking production. I built seven internal tools for the same firm, including the board that tracks its 45 AI initiatives.

Supabase RLSReview gatesClaude Code

Frontier model work

FRO · 25Ongoing25

Top frontier AI lab · Ongoing

Training environments & agent grading

I grade a top frontier lab's coding agents, build training data, and build the environments its models train in. Those environments are clones of business apps. Each one resets to the same starting data before every attempt and is scored on whether the model actually did the task. I once traced a bug where resetting one app wiped every other app's data in the shared database, down to one missing scoping wrapper. The lab goes unnamed by agreement.

RL environmentsAgent gradingTraining data

About

Production data at a payments company taught me what "it just has to work" means.

I interned at PhonePe, India's largest payments platform, working on lending analytics on production data, where a wrong number is someone's money. I turned down the return offer to build on my own.

I trained at Delhi Technological University and now run a small AI automation studio. Four of us have delivered for more than 20 businesses. I do the building. I take a real business problem, design the agentic system around it, ship it into production, watch it break, and fix it until it holds under load. No prototypes handed off to disappear.

On the side I work for a top frontier AI lab, grading its coding agents and building the environments its models train in. That keeps me close to what these systems can actually do. What I ship reflects where the models are today.

03 · How I work

Shipped means live and verified.

Start from the money

The brief is never the tech. It is the hour someone wastes or the deal that leaks. I automate the whole loop, then measure against that.

Own the outcome

I do not hand back a prototype and vanish. I run it in production, watch it break under real load, and fix it until it holds.

Build with the frontier

I grade a frontier lab's coding agents and build the environments its models train in, so I build for what this generation can actually do today.

Small team, real leverage

Four people and a software factory of coding agents. That is how we have delivered for more than 20 businesses.

Stack

The tools I reach for. The list moves as the models do.

Python TypeScript React Next.js Node.js Claude Gemini n8n Make Supabase Postgres Railway Stripe HubSpot Databricks ElevenLabs

Also working across VAPI, Retell, LiveKit, Deepgram, GoHighLevel, Attio, Instantly, and Apify.

05 · Let's build

Have a workflow worth automating? Let's talk.

The fastest path is a short call. Bring the problem. I will tell you honestly whether it is worth building and exactly how I would ship it.

Book a 30-minute call Email me

rahul@arraitech.com +91 77018 99550 LinkedIn