Voice agents
Outbound calling in your customer’s language, with outcomes classified and follow-ups booked automatically.
AI automation engineer / Pune, India
You describe the boring part.I ship the agent that does it.
I build voice agents, scraping pipelines, LLM apps and workflow automation for founders in the UK and Poland. Twelve of them build this page live on your first visit, and I check the work.
I am a Computer Science undergraduate in Pune who runs an automation studio. Founders in the UK and Poland tell me what their teams do by hand. I scope it, build it, ship it and hand over the docs.
01 / Work
Four jobs, shipped.
02 / What I build
Four things I get hired to build.
Scraping pipelines
Playwright scrapers that recover, dedupe and prioritise, then hand clean records to whatever comes next.
LLM-powered apps
React Native apps with a real backend behind them: caching, OCR, large databases and an assistant that knows the data.
Workflow automation
CRM, booking, content and sales tools wired together, so the manual steps stop.
03 / Proof
The numbers I can stand behind.
$14,000+Delivered revenueASTITWA AI
200–300Calls a dayHefes Voice
<200 msLatency, down from 3 sGoodFor
1,000+Listings scraped a dayHefes Voice
95%+Barcode and OCR accuracyGoodFor
99.9%Uptime at 100+ concurrent scansGoodFor
04 / How a job runs
Plain words in. Production out.
- 01
Scope
You explain the job in plain language. I turn it into a spec.
- 02
Architect
The backend comes first, before any prompt gets written.
- 03
Integrate
Claude or GPT-4, wired into your real tools: CRMs, telephony, booking systems.
- 04
Deploy
Docker and GCP. It runs in production, not in a demo.
- 05
Hand off
Documented and maintainable, so you are never stuck with me.
05 / Where
Where I have been working.
2025 to nowASTITWA AIFounder and AI Automation EngineerRemote, clients in the UK and Poland
Aug 2025 to nowHefesStrategic Technology PartnerPoland, remote
May to Jul 2026Go LegalSoftware and AI Engineering Intern (paid)London, UK, remote
Expected 2028MIT ADT UniversityB.Tech, Computer Science and EngineeringPune, India
06 / Stack
What is on the bench.
AI and LLM
- Anthropic Claude API
- OpenAI GPT-4 API
- Prompt engineering
- Agentic workflows
- AI voice agents
- NLP, OCR, computer vision
- Speech-to-text
- AI video generation
Automation
- Playwright
- Selenium
- BeautifulSoup
- Twilio / VoIP
- Tavily API
- Data pipelines
- CRM and booking integrations
- SEO automation
Backend and data
- Node.js
- Express.js
- REST APIs
- PostgreSQL
- Redis
- JWT authentication
- Docker
- Google Cloud Platform
Mobile
- React Native
- Expo
Languages
- Python
- JavaScript
- TypeScript
- Java
- SQL
- HTML and CSS
- Git and GitHub
07 / Your turn
So, what is your boring part?
Tell me what your team does by hand, again and again. I will tell you honestly whether an agent should be doing it.
