My JSON Resume

resume.json

{
  "basics": {
    "name": "Zac Saber",
    "label": "Frontend + ML Engineer",
    "image": "",
    "email": "[email protected]",
    "phone": "+61 411 895 085",
    "url": "[email protected]",
    "summary": "Hi, I'm Zac - ex-Atlassian, I'm currently doing ARENA, and finishing up a Frontend + ML role at Strong Compute. I've been hacking with machine learning since high school, when I built an NLP movie recommender with Scikit-Learn. But it was red-teaming GPT-4-0314 that made me fall in love (at the risk of sounding like a nerd...)",
    "location": {
      "address": "",
      "postalCode": "",
      "city": "Sydney",
      "countryCode": "AU",
      "region": "New South Wales"
    },
    "profiles": [
      {
        "network": "LinkedIn",
        "username": "z-saber",
        "url": "https://linkedin.com/in/z-saber"
      },
      {
        "network": "GitHub",
        "username": "zazer0",
        "url": "https://github.com/zazer0/"
      }
    ]
  },
  "work": [
    {
      "name": "Strong Compute",
      "position": "Frontend + ML Engineer",
      "url": "",
      "startDate": "2024-08",
      "endDate": "2025-04",
      "summary": "YC-funded; in SF",
      "highlights": [
        "Assisted AI Safety researchers in Berkeley (+ globally) with ML infrastructure challenges",
        "Ran monthly ML Training hackathons for researchers across Sydney + SF to improve accessibility of the field",
        "Engineered reactive UI to enable user-friendly cross-cloud ML job management. Used Svelte + TypeScript",
        "Architected + implemented scalable backend to handle concurrent usage with complex state. Also: e2e + unit tests"
      ]
    },
    {
      "name": "Various AI Startups",
      "position": "SysAdmin + DevOps Engineer",
      "startDate": "2023-12",
      "endDate": "2024-07",
      "highlights": [
        "Engineered initial PoC; up to OAuth2 in refined prod MVP. Azure vision model deployment to automate Xero invoicing",
        "Databases on GCP; highly-available Kubernetes clusters. Load-balanced backend across 3 sites; 99.9% uptime",
        "ETL pipelines to clean, chunk, embed & annotate data. Complex multi-modal LLM chatbot workflows; using Graph RAG"
      ]
    },
    {
      "name": "Atlassian",
      "position": "Product Security Engineer",
      "startDate": "2023-11",
      "endDate": "2024-02",
      "summary": "Internship",
      "highlights": [
        "Explained technical concepts to non-technical stakeholders. Created fully local secret scanner with <1 second runtime",
        "Learnt Go scripting and best deployment practices. Also: Security Reviews, microservice deployment, blogging"
      ]
    },
    {
      "name": "UNSW CSE Faculty",
      "position": "Casual Academic Tutor",
      "startDate": "2022-05",
      "endDate": "2024-07",
      "highlights": [
        "Tutor for COMP1511; selected to tutor right after finishing it! Taught 2x24-student classes the fun of programming",
        "Then tutored COMP2041; Python + POSIX scripting. Taught unix filters, APIs, web scraping, hacker mindset!"
      ]
    },
    {
      "name": "WiseTech Global",
      "position": "Junior SOC Analyst",
      "startDate": "2023-02",
      "endDate": "2023-06",
      "summary": "Internship",
      "highlights": [
        "Created flexible Python script suite to automate Tenable reporting, solving key UI bottlenecks",
        "Reduced reporting from hours -> minutes on CLI (bottlenecked by server!). Also: alerts, CDR's, SecEng"
      ]
    }
  ],
  "education": [
    {
      "institution": "TARA (AI Safety ANZ)",
      "area": "ARENA",
      "studyType": "In Person",
      "startDate": "2025-03",
      "endDate": "current",
      "summary": "A 12-week facilitated cohort of the ARENA program. On Transformers, Mech Interp, Evals, Sparse AutoEncoders. Running in Sydney, Australia. Includes dedicated project component"
    },
    {
      "institution": "University of New South Wales",
      "area": "Computer Science / Arts",
      "studyType": "Double Degree",
      "startDate": "2022",
      "endDate": "2025",
      "summary": "Majors: Artificial Intelligence, Photography. Minor: Psychology"
    },
    {
      "institution": "BlueDot Impact",
      "area": "AI Safety Fundamentals",
      "startDate": "2024",
      "endDate": "2024",
      "summary": "Technical Alignment Stream. Project measured effective benefit gained from RoPE in modern transformers"
    }
  ],
  "awards": [
    {
      "title": "1st Place SXSW Hackathon",
      "date": "2024",
      "summary": "HuggingFace tool to suggest an optimal model. Given a use case, scrapes HuggingFace to suggest optimal models. With pros/cons, summaries, & a friendly UI!"
    },
    {
      "title": "1st Place IMC Hackerrank Competition",
      "date": "2024",
      "summary": "Solved algos under time pressure. My team of 3 placed 1st in Division B. Won against teams a year ahead"
    },
    {
      "title": "1st Place Checkmarx Secure Coding",
      "date": "2024",
      "summary": "Assessed vulnerable codebases quickly. Placed 1st against working PhD students. Recognised by UNSW SecEDU"
    }
  ],
  "skills": [
    {
      "name": "Primary Skills",
      "keywords": ["Python", "Shell", "LLM's", "JavaScript", "ML"]
    },
    {
      "name": "Secondary Skills",
      "keywords": ["Docker", "Cloud", "GCP", "Kubernetes", "Git"]
    },
    {
      "name": "Additional Skills",
      "keywords": ["ETL", "Vector DB's", "Go", "React"]
    }
  ],
  "projects": [
    {
      "name": "Automated Code Style Marker",
      "description": "Deployed local, private, fine-tuned model. Analysed input context of 10,000 lines; based on custom marking rubric. Output friendly JSON comments; line numbers as keys, referencing specifics. Playwright to automate marking UI",
      "highlights": ["Azure", "Playwright", "PromptFlow"]
    },
    {
      "name": "FyshGPT: Spear Phisher",
      "description": "Webapp to demonstrate dangers of LLM's; circumvented OpenAI moderation. Generated tailored spear-phishing emails. Customisable for any custom context, objective, target. Also offers identification of key red flags for a suspected phish. And, vibe refinement of an existing phish",
      "url": "https://github.com/zazer0/Fysh-GPT",
      "highlights": ["Langchain", "Prompt Engineering", "Evals"]
    },
    {
      "name": "Homelab Server",
      "description": "Learnt networking OSI model; deployed microservices with Cloudflare. HAProxy to load-balance across Pi's, VM's, & bare-metal. Kubernetes to containerise & deploy for better reliability. Learnt GCP and Azure along the way!",
      "highlights": ["Kubernetes", "rPi/Proxmox", "HAProxy"]
    }
  ],
  "volunteer": [
    {
      "organization": "AI Society UNSW (AISOC)",
      "position": "Education Subcommittee",
      "startDate": "2024-04",
      "endDate": "2024-12",
      "highlights": [
        "Make machine learning more friendly for newbies. Host workshops to upskill non-technical students",
        "Translate technical developments into digestible content. Create easy-to-follow guides (e.g., on recent papers)"
      ]
    },
    {
      "organization": "UNSW Security Society",
      "position": "SCONES Conference Director",
      "startDate": "2022-11",
      "endDate": "2023-12",
      "highlights": [
        "Coordinated SecSoc's introductory security conference. Beat previous year; had 500+ attendees & 12+ diverse talks",
        "Sponsor liaising, budget planning, lots of teamwork. Handled all aspects of speakers including outreach + accessiblity"
      ]
    }
  ]
}