Mathew H.

Key Skills

Software

No software listed

Top Subject Matter

Retrieval evaluation

OCR-powered document intelligence

and ranking validation workflows

Top Data Types

Document

Text

Top Task Types

Text Generation

Question Answering

Freelancer Overview

Large-Scale Marketplace Matching and Ranking System (NDA). Professional background includes roles such as AI/ML Engineer. Core strengths include Internal, Proprietary Tooling, and OpenAI APIs. AI-training focus includes data types such as Document and labeling workflows including Evaluation, Rating, and Text Generation.

Labeling Experience

HakRakyat - AI Legal Assistant and Contract Analysis Platform

DocumentText Generation

Built and deployed an AI-powered legal assistant that performs OCR-driven document ingestion followed by contract analysis. Implemented workflow logic for premium assistant capabilities and secure feature gating. Delivered an end-to-end legal tech SaaS experience combining document understanding with subscription-oriented deployment. • Designed OCR–chunking–AI analysis pipelines.• Integrated contract analysis into an AI legal assistant workflow.• Implemented secure feature gating for premium tiers.• Achieved 1,000+ paying users in the first month and 73% retention into month 2.

2024 - Present

Large-Scale Marketplace Matching and Ranking System (NDA)

Document

Developed evaluation and validation workflows for retrieval and ranking systems that support AI-driven document understanding. Built replay-based testing pipelines to assess ranking correctness and latency before deployment. Designed enterprise-oriented document intelligence architecture leveraging OCR-derived inputs and retrieval orchestration.• Reduced retrieval latency from over 5 seconds to under 0.1 seconds.• Designed modular retrieval and ranking components for scalable deployment.• Planned embedding and hybrid retrieval infrastructure for improved retrieval quality.• Implemented replay-based evaluation and ranking validation for iterative model improvement.

2024 - Present

Multimodal Financial Document Intelligence System

DocumentQuestion Answering

Designed multimodal document intelligence pipelines to extract and reason over financial reports using text, tables, and visual information. Built context-aware retrieval orchestration workflows to improve answer grounding from multimodal inputs. Explored multimodal extraction and structural reasoning approaches to support enterprise document understanding. • Built context-aware retrieval orchestration workflows.• Explored multimodal extraction and structure reasoning systems.• Designed enterprise-oriented document intelligence architecture.• Implemented retrieval-aware insight generation pipelines.

2023 - Present

Education

Work History