Business · Finance · AI Operations

Nicholas
Fernando

Nicholas Fernando is a results-driven Business, Finance, and AI Operations professional with 9+ years of experience across financial operations, tax administration, customer support analytics, and AI training projects.

Finance & Tax Accounts Receivable AI Evaluation Data Annotation RLHF
Email Me WhatsApp LinkedIn
Career at a Glance
9+
Years professional experience
3
Domains: Business, Finance, AI
55%
Collection rate improved
7K+
Payments processed / week
2+
AI annotation & analytics projects
Remote
Available globally
Semarang
Indonesia-based
About Me

Professional Profile

Based in Semarang, Indonesia, Nicholas Fernando has contributed to AI data annotation, Large Language Model (LLM) evaluation, financial operations, taxation, customer support analytics, and business process optimization across multiple projects.

With a Computer Science background, Nicholas bridges the analytical rigor of finance with the technical demands of modern AI workflows — equally effective as a finance specialist and an AI operations contributor.

Open to remote opportunities globally in AI training, data annotation, business analysis, and financial operations roles.

9+
Years work experience
3
Core expertise domains
30%
Discrepancy reduction in AR
2+
AI projects completed
Career Evolution

Why Finance → AI?

"After building a strong foundation in finance, taxation, and operational reporting, I expanded into AI training and evaluation projects. My analytical background enables me to assess data quality, identify patterns, and contribute to AI model improvement through structured evaluation and annotation workflows."
💰
Finance & TaxAccounts Receivable, VAT, Reconciliation
→
📊
Data & AnalyticsReporting, Pattern Recognition, Quality Control
→
🧠
AI OperationsLLM Evaluation, Annotation, RLHF
Areas of Expertise

3 Core Domains

🏢
Business Operations
  • Process Improvement
  • Customer Operations
  • Business Reporting
  • Business Analysis
  • Remote Collaboration
💼
Finance & Tax
  • Accounts Receivable
  • Tax Administration
  • Financial Reporting
  • Payment Reconciliation
  • Debt & Collection Management
🤖
AI Training & Evaluation
  • Data Annotation
  • Large Language Model (LLM) Evaluation
  • Prompt Assessment
  • Reinforcement Learning from Human Feedback (RLHF)
  • Quality Assurance
Work Experience

Career Timeline

2024 — Present · AI Projects
AI Training & Evaluation Specialist
Remote — Freelance / Independent
  • Evaluated AI responses using rubric-based scoring (clarity, completeness, actionability, tone, accuracy)
  • Performed A/B ranking of model outputs with written justification and improvement notes
  • Detected and corrected factual inaccuracies and hallucinations in AI-generated content
  • Designed structured prompts with role, constraints, and output formatting for multiple use cases
  • Prepared structured datasets for Large Language Model (LLM) training workflows
LLM EvaluationData AnnotationPrompt EngineeringRLHF
2024 — 2025
Tax Staff
PT Berill Jaya Sejahtera
  • Managed VAT reporting, tax reconciliations, and full compliance processes
  • Improved collection rate by 55% through billing system optimization
  • Generated financial reporting for internal and external stakeholders
VAT ReportingTax ComplianceReconciliation
2021 — 2023
Finance AR Officer
Kawan Lama Group
  • Processed 7,000+ incoming payments weekly with high accuracy
  • Reduced payment discrepancies by 30% through systematic reconciliation
  • Improved claim resolution efficiency by 40%
Accounts ReceivableReconciliationClaim Resolution
2016 — 2021
Finance AR Staff
CV Sinar Anugerah
  • Managed full receivables cycle and customer debt analysis for 50+ client accounts
  • Identified growth opportunities that increased revenue by 25%
  • Maintained accurate financial records and generated monthly performance reports
Receivables ManagementDebt AnalysisRevenue Growth
Tools & Software

What I Work With

🤖 AI Tools
🤖 ChatGPT
🧠 Claude AI
🔍 Gemini
🌐 Perplexity
📊 Data & Analytics
📈 Microsoft Excel
📊 Google Sheets
📉 Looker Studio
⚙️ Productivity & Business
📝 Notion
💬 Slack
🔗 Zapier
📦 Google Workspace
🎨 Design & Content
🎨 Canva
🎬 CapCut
✍️ Writing & Docs
✍️ Grammarly
📄 Google Docs
Portfolio Projects

Work Samples & Case Studies

⭐ Featured Case Study · AI Evaluation
Customer Support Data Annotation & Quality Review
Role: AI Evaluator & Annotator
Type: AI Training Dataset
Tools: Google Sheets, Rubric Framework
82
Avg. Score /100
🔍 Problem Statement

AI-generated customer support responses often lack consistency in tone, accuracy, and completeness — creating a poor user experience and reducing trust in automated support systems. Without structured human evaluation, these issues go undetected before deployment.

👤 My Role & What I Did
  • Designed a multi-criteria evaluation rubric covering accuracy, relevance, completeness, and safety
  • Reviewed and labeled 100+ customer support conversations for sentiment and urgency
  • Scored each AI response against the rubric and documented reasoning for each verdict
  • Identified recurring failure patterns across response types
  • Prepared a structured dataset ready for RLHF training pipeline ingestion
💡 Key Insights
  • Most failures occurred in safety (10%) and completeness (20%) criteria
  • Responses scored higher on relevance but frequently lacked specific action steps
  • Urgent queries were often answered with generic, low-urgency language
📋 Evaluation Rubric
CriteriaWeightScore
Accuracy40%
85
Relevance30%
88
Completeness20%
72
Safety10%
78
📝 Example AI Output Evaluated
"Your order has been received. Please wait 3–5 business days for delivery. Contact us if you have questions."
Analysis (Score: 82/100): The response is accurate and relevant but lacks specific tracking information and an empathetic tone for a delayed order complaint. The completeness score is reduced due to missing action steps and no reference to escalation options.
Recommendation: Add a tracking link, acknowledge the inconvenience explicitly, and include a direct escalation path (e.g., live chat or phone). This would bring completeness from 72 → 90+.
📈 Results & Impact

The annotated dataset provided a structured, rubric-aligned training signal for the AI model. By categorizing failure types and documenting reasoning, the evaluation output enabled the model fine-tuning team to prioritize improvements in completeness and safety — the two lowest-scoring criteria. The structured format also reduced review time by making each verdict immediately actionable.

🔄 Improvement Recommendations
  • Expand rubric to include Tone & Empathy as a standalone criterion for support-specific evaluation
  • Add urgency classification labels to improve response prioritization in training data
  • Introduce inter-annotator agreement checks to reduce subjectivity in borderline cases
Data Annotation LLM Evaluation Rubric Design RLHF Quality Assurance Google Sheets
View Dataset
🏨
Data Analytics · Finance
Hotel Booking Analytics

Problem Statement

Hotel management lacked visibility into booking patterns and cancellation drivers, making it difficult to optimize pricing and occupancy strategies.

What I Did

Analyzed booking data to surface revenue trends, seasonal patterns, cancellation rates, and customer segmentation insights using structured spreadsheet analysis.

Key Insight

Weekend bookings showed 2× lower cancellation rates — suggesting targeted promotions could shift revenue mix toward higher-retention segments.
Data AnalysisRevenue TrendsBooking PatternsSegmentation
Google Sheets · Analytics View Project
💡
AI Evaluation · Methodology
My Evaluation Framework

How I Think

When evaluating an AI response, I ask: Is it accurate? Is it complete? Would it actually help the user? I look beyond surface correctness to assess whether the response creates trust or confusion.

Decision Process

I assign scores per criterion, document specific failure reasons (not just a low score), and always conclude with a concrete improvement recommendation — not just a verdict.
Critical ThinkingAnalytical JudgmentRubric Design
Methodology · Process See Case Study ↑
Education

Academic Background

🎓
Bachelor of Computer Science
Semarang University
A strong technical foundation bridging analytical thinking with AI and technology roles.
Get In Touch

Let's Work Together

Open to remote opportunities — AI evaluation, prompt engineering, data annotation, finance consulting, or business operations roles.

nicholassfern@gmail.com LinkedIn