Vlad Bulaev · AI & Data Engineer · Melbourne

I build AI systems end to end.

Here are five examples of my work, each built and launched solo, from raw data to production. Looking for an entry-level role in AI, data engineering or analytics in Melbourne.

AI EngineeringData EngineeringAI agentsSoftware Development
Portrait of Vlad

I’m Vlad, a Bachelor of ICT Engineering graduate specialising in Data Engineering and Analytics, with hands-on experience developing software, working with data and building practical technology solutions.

  • Full working rightsin Australia
  • English B2upper-intermediate
  • Office or remoteready for either

Portfolio

Selected projects.

Real problems, what I built and what it delivered.

AI sales automation

AI that makes cold calls and qualifies the prospects.

In plain words: an AI assistant phones your potential customers, talks to them like a person and hands you only the ones who are interested.

  • 2,000+calls made by the AI
  • 1permanent job contract won
  • 1family moved to Australia

ResultThe campaign led to a permanent marketing role for my wife with an Australian employer, who then sponsored our permanent residency. That is how we came to live in Australia.

What it can do

  • Sounds naturalA human-like voice that holds a real conversation.
  • Answers fastNo awkward pauses or delays.
  • Asks you when unsureIf it doesn’t know the answer, it sends you the question in Telegram with reply buttons.
  • Live transcriptionYou see the whole conversation as it happens.
  • Control by chatType a message and the AI says it to the person on the call.
  • End-call buttonYou can stop the call at any moment.

A real example

Listen to a real callA recording of the AI assistant talking to a prospect.
Read the full story

I started by scraping targeted data to build a high-intent list of phone numbers. Instead of generic lists, I built targeted datasets of leads that closely match the ideal customer profile. Then AI made the cold calls, introduced the services and managed the conversation in real time.

If a prospect expressed interest, the system automatically qualified the lead, generated a full transcription of the call and delivered the data directly to me. The voice assistant holds a natural, human-like dialogue: it handles objections, explains value propositions and extracts qualification data as the call happens.

    Built withVapiTwilioCloudflareClaudeOpenClawElevenLabs

    My roleBuilt solo: lead scraping, the calling agent, conversation logic, transcription and delivery of results.

    Market research & data engineering

    A clickable map for an expansion decision.

    In plain words: a company wanted to open an office in Australia with no data to go on. I gathered it and turned it into a map that shows where customers are many and competitors are few.

    ResultThe client opened a branch in Australia, a decision based in part on this analysis.

    Read the full story

    My client wanted to open an office in Australia but had zero data to support the decision. There was no internal research, no competitor map and no demographic breakdown. They were essentially flying blind into an expansion. I took total ownership of the discovery phase and acted as a one-person data engineering team to scrape, clean and visualize the Australian market from scratch.

    Since no dataset existed, I went to the source. I engineered scrapers to pull demographic data from the Australian Bureau of Statistics (ABS) and legal and competitor data from the Migration Agents Registration Authority (MARA). I took data from completely different worlds, government population tables and unstructured web directories, and built a unified relational database.

      Customer demand
      Competitor density
      Opportunity

      Demo with illustrative index values (0 to 100), not the client’s real figures. Tap a state and switch the layer.

      Built withPythonSeleniumBeautifulSoupPandasHTML / CSS

      My roleBuilt solo: the scrapers, data cleaning and merging, the database and the interactive map.

      Full-stack AI SaaS · in production

      An AI platform that runs a migration agency.

      app.sivavisa.com ↗In use by real clients

      In plain words: a visa agency worked in spreadsheets and endless email threads. I built one system that reads client documents, checks them against visa rules, shows what is missing and keeps clients, employers and agents in one place.

      • 60+database tables
      • 30+API modules
      • 16+visa types covered
      • 12connected services
      • 7background workers
      • 130new leads captured

      ResultThe platform has already captured 130 new leads through its built-in tools. Over the same period the revenue of the agency’s Australian branch grew 4×, helped in part by the platform.

      Read the full story

      Migration agencies relied on spreadsheets, email threads and manual document review, leaving clients in the dark and agents overwhelmed with repetitive work. A single visa application can mean checking 10+ criteria, cross-referencing 36 months of payslips, validating English scores and coordinating between employer, client and agent across months of back-and-forth.

      The SivaVisa AI Migration Platform replaces that with one workflow, from the first lead to the lodgement pack: an agent portal, a client portal and an employer flow on a single backend. Every AI decision is logged, and every AI-written email waits for a human agent to approve it. Built and shipped solo.

      AI modulesBusiness toolsPlatform
      One platform

      1. A document:Upload
      2. Read (OCR)
      3. Classify
      4. Match to rules
      5. Find gaps
      6. Agent approves
      Built withPythonFastAPIReactPostgreSQL + pgvectorDockerClaudeGoogle Document AIAWS S3Stripe

      My roleDesigned, built and deployed solo: database, backend, two portals, AI modules and integrations.

      AI content automation

      A content factory that writes in your voice.

      In plain words: you drop a one-line idea into a private chat. A few minutes later a finished post, written in your own style, is waiting for your approval.

      • 200past posts studied to learn the style
      • 5 secto pick up a new idea
      • 1–3 minto a finished draft

      ResultMore than 100 social media posts generated so far, saving many hours of copywriting.

      What it can do

      • Writes in your styleLearns from your past posts and copies the way you write.
      • Works on its ownYou drop an idea, the agent writes the post.
      • Rewrites on requestDon’t like a part? Say “make it shorter” and a new version arrives in a minute.
      • Always runningStarts automatically with the computer or server and restarts itself if it crashes.

        A real example

        Ideas channelYou

        What is Victoria state famous about?

        3 minlater →
        Drafts channelAI agent

        What is Victoria actually famous for? 🇦🇺

        Outside Australia you’ll hear “coffee” and “the Australian Open.” Both true. But for anyone planning to move, Victoria rests on three numbers that matter more than a good flat white.

        1️⃣ Most liveable city in the country. In the 2026 EIU index, Melbourne ranked 3rd in the world, behind only Copenhagen and Vienna, scoring 97/100, with perfect marks for healthcare and education.

        2️⃣ Fastest-growing capital. Melbourne holds ~5.44M people, about 200,000 behind Sydney, but grew +105,000 last year vs Sydney’s +75,200. On trend, it overtakes Sydney as Australia’s largest city around 2030–2031.

        3️⃣ A real migration pathway. For 2025–26 Victoria got 3,400 nomination places: 2,700 for visa 190, 700 for 491. ROIs closed 28 April 2026, but this is what turns “I like Melbourne” into PR.

        The laneways, the AFL, the Great Ocean Road: that’s why people visit. The liveability and the nomination program: that’s why they stay.

        Built withPythonClaude CodeTelegram Bot APITelethon

        My roleBuilt solo: the engine, the Telegram integration, style learning and auto-start.

        AI customer assistant

        An AI assistant that answers your customers in Telegram.

        In plain words: customers write to your business at any hour. The assistant replies like a person, knows your prices and rules, and gently leads the conversation to a sale.

        ResultThe client now needs fewer moderators in their business chat, which cut their support costs.

        What it can do

        • Replies like a personNatural, friendly answers, not canned scripts.
        • Works 24/7No customer waits until morning.
        • Knows your businessPrices, FAQ, products and rules, all from your knowledge base.
        • Speaks their languageAnswers in the language the customer writes in.
        • Leads to the purchaseDoesn’t just answer, it guides people to the next step.
        • Follows upIf someone didn’t buy, it sends a personal message a day later, based on the conversation.

        Example dialogue for an invented business. Pick a situation to see how the assistant responds.

        Built withClaude CodeOpenClawTelegramKnowledge base

        My roleBuilt solo: the agent, the knowledge base, conversation logic and follow-ups.

        What I do

        Software, data and
        models that work
        in the real world.

        AI Engineering

        AI assistants, agents and document intelligence built on modern language models.

        Data Engineering and Analytics

        Scraping, cleaning and analysing data end to end, so it is structured, ready to use and answers real business questions.

        Software Development

        Hands-on experience building practical technology solutions.

        What’s next

        Open to entry-level roles in Melbourne.

        RolesJunior AI Engineer, Junior Data Engineer, Python Developer, Data Analyst, AI Automation Developer.
        Work rightsFull working rights in Australia.
        FormatOffice, hybrid or remote. Based in Melbourne.

        Let’s free people from the routine that can be automated, so they can focus on what truly matters.

        Let’s talk.

        Based in Melbourne. Open to entry-level roles in AI, data engineering and analytics.

        smartbee974@gmail.com
        0461 280 959