Krisztian Simon - AI Trainer & Evaluation Specialist

Recommended expert

Available

Oradea, Romania

Check rate

Experience

Apr 2026 - Aug 2026

5 months

Freelance Evaluator & Human Data Reviewer

Independent AI Training & Evaluation Projects

Evaluate AI-generated outputs against detailed rubrics covering relevance, accuracy, groundedness, safety, naturalness and task completion.
Review simulated storefront and support-agent conversations for groundedness, policy compliance, naturalness and task completion.
Compare AI-generated image and video outputs for prompt alignment, visual consistency, artifacts, realism and temporal coherence.
Apply changing calibration guidance and edge-case policies consistently across high-volume evaluation workflows.
Write concise, evidence-based scoring rationales that separate observed facts from assumptions.
Handle e-commerce product classification, catalog QA, search quality evaluation, category tagging and attribute extraction for structured datasets.

Jan 2026 - Present

8 months

Sole Developer & Project Lead

ZeroBounce

Self-initiated and building solo: a native Salesforce integration approved by the executive team to replace an unreliable third-party connector.
Writing the full solution in Apex and JavaScript, deploying and testing through Salesforce CLI, and validating behavior manually against production traffic.
Building it alongside a full-time lead role by driving coding agents (Claude Code, Codex) through design, implementation and test, reviewing every change and verifying against live API behavior.
Presented architecture, packaging and AppExchange distribution plans to engineering and executive leadership.

Jan 2026 - Present

8 months

Adversarial VQA Benchmark Creation

micro1

Create image-question-answer benchmark tasks (211 submitted to date) designed to make current multimodal models fail on dense charts, diagrams and detailed photographs.
Write questions that require several reasoning steps, with exact ground-truth answers a reviewer can verify directly against the source image.
Design benchmark items that target multimodal reasoning weaknesses while keeping answers objective, reproducible and verifiable.
Apply reviewer feedback and calibration guidance to raise task difficulty and remove ambiguity. Every task is individually reviewed before acceptance.
Source public-domain and CC0 imagery so every benchmark item is clean and reproducible.

May 2023 - Present

3 years 4 months

Technical Enablement Lead (Technical Support Manager)

ZeroBounce

Designed the support domain for an internal AI agent assessment system: simulated customer conversations, technical scenarios and scoring rubrics, with human final approval on every result.
Administer Zendesk and drive AI Copilot tuning through workflow review, agent feedback, response-quality analysis and recurring case patterns.
Lead technical enablement for a 30+ person Support and Sales team and directly manage one enablement specialist.
Tier-3 escalation owner for REST API, CRM integration, SMTP, authentication and DNS failures. Cut escalation volume 28% by converting the recurring cases into SOPs and targeted training.
Built and maintain a 210+ page knowledge base covering email validation, REST API behavior, status and substatus logic, and CRM integrations, alongside 15+ SOPs and decision-tree workflows.
Rebuilt new-hire onboarding as a scenario-based curriculum around API authentication, JSON payloads and live technical cases, cutting new-hire ramp from six weeks to three.
Monitor API traffic in Cloudflare, investigate abusive usage patterns, coordinate customer outreach, and refine traffic-control rules.

Aug 2022 - May 2023

10 months

Technical Support Engineer

SendGrid (via IntelligentBee)

Supported SMB through enterprise customers on SMTP, REST APIs, JSON payloads, authentication, SPF/DKIM/DMARC, suppression lists, bounce classification and IP warmup.
Traced delivery behavior in Splunk, Snowflake and Postman, then explained root cause directly to customers.
Handled 13+ tickets and 25+ live chats daily within SLA. Promoted to TSE+ for technical performance and for mentoring newer engineers.

Dec 2019 - Aug 2022

2 years 9 months

Oradea, Romania

Technical Support Engineer

DIGI (RCS-RDS)

Handled B2B and residential network escalations across MPLS, BGP, broadband, mobile, GPON/PON and VLAN environments under SLA-driven conditions.
Monitored network status across a three-county area on overnight operations and coordinated field teams to restore connectivity.

Jun 2017 - Dec 2019

2 years 7 months

Oradea, Romania

Customer Support Agent

DIGI (RCS-RDS)

Handled 70+ daily support inquiries covering connectivity, router configuration and account issues, and helped onboard new agents.

Industry Experience

See where this freelancer has spent most of their professional time.

Experienced in Telecommunication, Information Technology, and Retail.

Telecommunication

Information Technology

Retail

Business Area Experience

See which departments and functions this freelancer has contributed to most.

Experienced in Customer Service, Information Technology, Operations, Product Development, Quality Assurance, and Project Management.

Customer Service

Information Technology

Operations

Product Development

Quality Assurance

Project Management

Summary

AI trainer and evaluation specialist with a 9-year technical background in SaaS support, Tier-3 escalations, API operations and enablement. Builds adversarial VQA benchmark tasks for multimodal model testing and evaluates LLM outputs against detailed rubrics across text, image, audio and video. At ZeroBounce, owns the support-domain design of an internal AI agent assessment system: simulated conversations, technical scenarios, scoring guidance and guardrails, with human final approval. Years of root-cause analysis and technical writing underpin the evaluation work: precise rationales, consistent calibration, and answers a reviewer can verify independently.

Skills

Ai Evaluation & Human-In-The-Loop Systems: Llm Output Review, Adversarial Vqa, Rlhf/Sft Tasks, Response Ranking, Rubric-Based Scoring, Hallucination Detection, Groundedness, Factuality, Safety, Scenario Design, Guardrails, Human Final Approval
Human Data & Annotation: Visual Question Answering, Data Labeling, Product Classification, Catalog Qa, Category Tagging, Attribute Extraction, Search Quality Evaluation, Tts And Audio Review, Image And Video Prompt Alignment
Agent Assessment & Quality: Simulated Support Conversations, Technical Knowledge Evaluation, Scoring Guidance, Calibration, Qa Feedback, Scenario-Based Assessment
Saas & Api Systems: Rest Apis, Json, Webhooks, Authentication Flows, Postman, Cloudflare, Salesforce (Apex, Salesforce Cli), Hubspot, Zendesk, Jira, Splunk, Snowflake
Email Infrastructure: Smtp, Dns, Spf, Dkim, Dmarc, Bimi, Ip And Domain Warmup, Bounce Analysis, Inbox Placement, Suppression Lists

Languages

Romanian

Native

English

Advanced

Hungarian

Elementary

Education

Sep 2011 - Jun 2015

Simion Bărnuțiu National College

Romanian Baccalaureate · Philology · Șimleu Silvaniei, Romania

Certifications & licenses

HubSpot Sales Hub Software Certified

HubSpot Service Hub Software Certified

Statistics

Experience

Total positions 7

Experience in Telecommunication 5 y

Avg length 1 y 6 m

Longest experience 3 y 3 m

Global Experience

Countries worked in 1 (Romania)

Primary country Romania

Expertise

Recent roles Freelance Evaluator & Human Data Reviewer, Sole Developer & Project Lead, Adversarial VQA Benchmark Creation

Main industries Telecommunication, Information Technology, Retail

Main business areas Customer Service, Information Technology, Operations

Qualifications

Certifications earned 2

Profile

Created

August 2026

Need a freelancer? Find your match in seconds.

Try FRATCH GPT

Frequently asked questions

Have questions? Find more information here.

Krisztian is based in Oradea, Romania and prefers 100% remote projects.

Krisztian speaks the following languages: Romanian (Native), English (Advanced), Hungarian (Elementary).

Krisztian has at least 9 years of experience. During this time, Krisztian has worked in at least 6 different roles and for 5 different companies. The average length of individual experience is 1 year and 4 months. Note that Krisztian may not have shared all experience and actually has more experience.

Based on recent experience, Krisztian would be well-suited for roles such as: Freelance Evaluator & Human Data Reviewer, Sole Developer & Project Lead, Adversarial VQA Benchmark Creation.

Krisztian's most recent position is Freelance Evaluator & Human Data Reviewer at Independent AI Training & Evaluation Projects.

In recent years, Krisztian has worked for Independent AI Training & Evaluation Projects, ZeroBounce, micro1, SendGrid (via IntelligentBee), and DIGI (RCS-RDS).

Krisztian is most experienced in industries like Telecommunication, Information Technology, and Retail.

Krisztian is most experienced in business areas like Customer Service, Information Technology, and Operations. Krisztian also has some experience in Product Development, Quality Assurance, and Project Management.

Krisztian has recently worked in industries like Information Technology, Telecommunication, and Retail.

Krisztian has recently worked in business areas like Customer Service, Information Technology, and Operations.

Krisztian attended Simion Bărnuțiu National College for Philology.

Krisztian has 2 certificates. These include: HubSpot Sales Hub Software Certified and HubSpot Service Hub Software Certified.

Krisztian is immediately available full-time for suitable projects.

Daily Rate Distribution

0 2 4 6 8

<€400 €400-440 €600+

The rates shown represent the typical market range for freelancers in this position based on recent contracts on our platform.

Actual rates may vary depending on seniority level, experience, skill specialization, project complexity, and engagement length.

Average rates for similar positions

Rates are based on recent contracts and do not include FRATCH margin.

800

600

400

200

Daily rate avg. 572 €

The average daily rate is the mean of all daily rates from recent contracts of comparable freelancers on our platform.

800

600

400

200

Median rate 600 €

The median daily rate is the middle value of all daily rates — half of comparable freelancers charge less, half charge more. Unlike the average, it is barely affected by outliers.

Actual rates may vary depending on seniority level, experience, skill specialization, project complexity, and engagement length.