Fisor Builder

Research/data pipeline module converting natural language questions into structured datasets.

Fisor

Overview

Fisor Builder is the ingestion and processing engine for Fisor Analytics. It combines LLM reasoning (Cohere) with real-time web search (Perplexity) to iteratively generate sub-queries and build validated, structured datasets from open-ended questions.

Key Features

  • Iterative Search Loop: Query -> Search -> Parse
  • Sub-query Generation: Uses Cohere Command R+
  • Real-time Data: Perplexity API integration
  • Validation: Custom confidence and schema validators
  • Performance: JSON/Redis caching and Prometheus metrics

Tools & Technologies

Python
Cohere API
Perplexity API
Docker
Redis

Links & Resources

Dedicated and hardworking Computer Science student with a passion for software development, committed to delivering outstanding results and exceeding expectations.

© 2024-2025 Enes Yesil. All Rights Reserved.