JJ657

Data & retrieval pipeline builder. Researching local vector database tradeoffs, text chunking heuristics, embedding token economics, and low-latency RAG.

Madrid, Spain Joined on Sep 5, 2026

Work

Vector Data & Retrieval Engineer

JJ657

Oct 18

RAG Context Compression: Headroom Cuts Retrieval Token Costs by 57% Without Accuracy Loss

#python #ai #datascience #opensource

4 min read

JJ657

Oct 17

Graphify: Deterministic Code Knowledge Graphs vs. Vector RAG Token Waste

#python #ai #datascience #opensource

3 min read

JJ657

Oct 16

PageIndex RAG: Stop Paying Twice for Overlapping Evidence

#python #ai #datascience #opensource

4 min read

JJ657

Oct 5

Deep Dive: rohitg00/ai-engineering-from-scratch Architecture

#python #ai #datascience #opensource

4 min read

JJ657

Sep 28

Making RAG Answers Shorter Without Breaking Production: An Output Contract Around i-have-adhd

#python #ai #datascience #opensource

4 min read

JJ657

Sep 21

Solving Context Rot: Integrating Tencent/WeKnora for Cache-Aligned RAG and Token Control

#python #ai #datascience #opensource

3 min read

JJ657

Sep 15

Taming Agent Context Inflation: Integrating OpenViking for Cache-Aligned RAG and Token Optimization

#python #ai #datascience #opensource

3 min read

JJ657

Sep 14

Making UltraRAG retrieval cheaper with stable prompt prefixes

#python #ai #datascience #opensource

3 min read

JJ657

Sep 8

Turning public-apis into a Cache-Friendly RAG Retrieval Layer

#python #ai #datascience #opensource

5 min read

DEV Community

JJ657

Badges

Writing Debut

Skills/Languages

Currently learning

Currently hacking on

Available for

RAG Context Compression: Headroom Cuts Retrieval Token Costs by 57% Without Accuracy Loss

Graphify: Deterministic Code Knowledge Graphs vs. Vector RAG Token Waste

PageIndex RAG: Stop Paying Twice for Overlapping Evidence

Deep Dive: rohitg00/ai-engineering-from-scratch Architecture

Making RAG Answers Shorter Without Breaking Production: An Output Contract Around i-have-adhd

Solving Context Rot: Integrating Tencent/WeKnora for Cache-Aligned RAG and Token Control

Taming Agent Context Inflation: Integrating OpenViking for Cache-Aligned RAG and Token Optimization

Making UltraRAG retrieval cheaper with stable prompt prefixes

Turning public-apis into a Cache-Friendly RAG Retrieval Layer