Kennedy C.

Key Skills

Software

Don't disclose

Other

Top Subject Matter

Enterprise document RAG (PDF ingestion, semantic vector search, retrieval-augmented QA)

LLM prompt workflows for developer tooling (error-to-fix)

AI risk assessment and risk-level classification for real-time fraud detection

Top Data Types

Document

Image

Top Task Types

Question Answering

Prompt + Response Writing (SFT)

Classification

Freelancer Overview

Azeru — Enterprise RAG Platform (PDF semantic search and QA). Brings 1+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose and Other. Education includes Bachelor of Technology, Federal University of Technology, Owerri (FUTO). AI-training focus includes data types such as Document, Computer Code, and Programming and labeling workflows including Question Answering, Prompt + Response Writing (SFT), and Classification.

Labeling Experience

Asguard — AI Fraud Detection System (classification + explanatory summaries)

Don't discloseClassification

Implemented a real-time fraud detection system that classifies transactions into risk levels and generates natural-language explanations. Combined rule-based scoring with an AI risk assessment engine to flag suspicious events in high-volume streams. Produced summaries for flagged events to support operational review and decision-making.• Built a scalable Go fraud detection backend with sub-second latency goals.• Implemented risk classification (Low/Med/High) based on scoring logic.• Generated AI natural-language summaries for flagged transactions.• Designed a real-time pipeline for transaction risk assessment.

2024 - 2024

Vpipe — AI CLI Debugger (multi-provider LLM prompt/response workflow)

OtherPrompt + Response Writing (SFT)

Developed an AI CLI debugger that transforms build errors into actionable fix suggestions using multi-provider LLM calls. Implemented preprocessing to sanitize sensitive patterns before sending prompts externally and truncation to preserve critical error context within token limits. Ensured the assistant produced consistent, structured guidance suitable for developer workflows.• Routed build errors through Groq/OpenAI/Anthropic for instant suggestions.• Built an auto-sanitization engine to mask API keys, IPs, and tokens before LLM requests.• Added smart truncation to respect provider token limits while retaining key context.• Designed prompt/response workflow for practical coding assistance outputs.

2024 - 2024

Azeru — Enterprise RAG Platform (PDF semantic search and QA)

Don't discloseDocumentQuestion Answering

Built an enterprise RAG pipeline that ingests PDFs into a semantic knowledge base for context-aware question answering. Implemented intelligent text chunking and retrieval using vector search to support citation-backed AI responses. Focused on optimizing query latency and relevance scoring for accurate, user-facing AI answers.• Ingested and indexed PDF content for semantic retrieval.• Designed chunking and conversational retrieval to answer private document questions.• Integrated Groq (Llama 3/4) for LLM-driven responses with relevant context.• Tuned relevance scoring and response generation to improve latency and answer quality.

2024 - 2024

Education

Work History