Skip to content

Data Scientist & Software Engineer

Hi there, I’mZsófia Németh

I'm currently pursuing a double degree in Computer Science @ELTE and Data Science in Business @Corvinus, though with 3 years in Data Science and Web Development working at Forward College, my professional experience actually outdates my studies! Alongside my degree, I'm the co-founder of Dropin, an aggregated booking platform in Hungary. I am also working on computer vision related research projects, both individual (NTP funded) and one in a research group with Harvard Medical School. In my free time, I love diving into Cybersecurity and Machine Learning (hoping to continue my studies in one or both :) ). I have lived in 5 different countries, and I can't wait to extend this list.

Download PDF CVDownload ATS text
Portrait of Zsófia Németh
nemethzsofia442@gmail.comLinkedIn

My history in bullet points

(Newest first)

  1. Education

    Data Science in Business

    Corvinus University of Budapest · Budapest, Hungary

    • Continues LSE studies in Data Science and Business Analytics through a Data Science in Business degree.
    • Conducts NTP scholarship-supported machine learning research on image recognition in data-scarce environments.

  2. Education

    Erasmus Exchange Studies

    University of Primorska · Koper, Slovenia

    • Broadened applied data science skills through Erasmus coursework in Data Programming, Data Mining and Machine Learning, Augmented Reality, and Adaptive Interactive Systems.

  3. Education

    Erasmus Exchange Studies

    Universidad Complutense de Madrid (UCM) · Madrid, Spain

    • Strengthened systems fundamentals through Erasmus coursework in Computer Organisation, Databases, Computer Networks, and Operating Systems.

  4. Education

    BASc Computer Science

    Eötvös Loránd University (ELTE) · Budapest, Hungary

    • Deepened core software-engineering foundations, by completing coursework in object-oriented design and systems programming.
    • Practised desktop application architecture, by building a Cinema Management System in C# and a Parking Lot Management System in Java.

  5. Work

    Data Scientist

    Forward College · Budapest, Hungary

    • Shipped a production multi-model AI workspace ("AI Hub") for a closed community of ~500 students and staff, by building a ~55,000-line TypeScript/Python platform across React, 14 Supabase Edge Functions, and 4 Python Cloud Run services.
    • Unified 3 LLM providers (OpenAI, Anthropic, Google Gemini) behind a single streaming interface with ~15 configurable tools and up to 10 tool cycles per request, by writing a provider-routing SSE orchestration layer with token-budgeted context management (80k-token ceiling reduced to ~30k).
    • Cut retrieval context passed to the main model by roughly 60–67%, by designing a multi-source RAG pipeline (OpenAI embeddings, Pinecone, Cohere reranking) that narrows 30 semantic candidates to 10–12 reranked chunks behind server-enforced per-user filters.
    • Verified sign-up reliability at 200 concurrent registrations with a 100% success rate (p95 8.75 s), by k6 load-testing the auth pipeline and tracing the 500-user failure mode to a 15-connection PostgreSQL pool.
    • Protected respondent privacy across multi-period 360° feedback research while keeping identities analytically consistent, by building a Python anonymisation pipeline with keyed HMAC-SHA256 pseudonyms, regex-based PII detection, and free-text name redaction.
    • Reduced transcription API payloads by up to ~97% for hour-long recordings (≈1 GB video → ≈29 MB audio), by preprocessing uploads with FFmpeg into mono 16 kHz MP3 before word-level Deepgram Nova-3 transcription.
    • Digitised qualitative admissions research built around four-section, ~20-minute interviews, by developing browser-based recording tools with Google Cloud Speech-to-Text transcription and automated transcript export.
    • Converted prospective-student interest into qualified CRM leads, by shipping a data-driven financial-aid eligibility simulator powered by four CSV rule datasets and integrated with the HubSpot Forms API.

  6. Work

    Website Developer

    Danon Jewellery LTD · Hungary

    • Launched the Hungarian webshop for a franchised jewellery brand serving an international customer base, by designing and implementing the storefront end to end.
    • Brought a catalogue of 200+ products online, by structuring, localising, and publishing the full product range for the new market.

  7. Work

    Mathematics Student Assistant

    Forward College · Lisbon, Portugal

    • Prepared incoming students for university-level mathematics, by designing and running a comprehensive summer bootcamp from scratch.
    • Made the material self-serve and reusable, by producing multimedia educational content and maintaining a dedicated Moodle learning platform.

  8. Work

    Student Ambassador

    Forward College · Lisbon, Portugal

    • Eased prospective students' transition into college life, by acting as their primary liaison and supporting them through engagement and induction-week activities.

  9. Education

    BSc Data Science and Business Analytics

    The London School of Economics and Political Science (LSE), via Forward College · Lisbon, Portugal

    • Combined an LSE data science and business analytics degree with four years of hands-on startup work, by studying through Forward College's practice-based campus model.
    • Applied coursework directly to production systems, by building the college's AI platform, research tooling, and data pipelines while studying.