Looking for a job

Salih Furkan Erik

@salihfurkanerik

AI Engineer at FloLabs R&D building FloBrain, an LLM orchestration core that connects wearable telemetry with robotic hardware.

Turkey

Message

At FloLabs R&D, I’m leading development of FloBrain, the LLM orchestration core of MoodChanger. I consolidated five fragmented data models into a stateful pipeline for real-time telemetry from more than 12 wearable types.

I designed a Control-Worker-Judge agent framework with sub-100ms sensor fusion, connecting cloud LLM reasoning to edge wearable hardware. I’m also building motor control abstraction layers that translate model outputs into actuator commands across three robotic platforms.

My work spans speech, vision, and open-source models. At Callmetric AI, I fine-tuned speech-to-text models with LoRA and developed mel-spectrogram analysis for emotionally ambiguous speech; my published models have attracted more than 6,000 combined downloads.

Experience

Work history, roles, and key accomplishments

Current

Embodied AI Engineer Intern, Team Lead

Current

FloLabs R&D

Feb 2026 - Present (9 months)

Leading development of FloBrain, the LLM orchestration core of MoodChanger, consolidating fragmented data models into a unified stateful pipeline. Designed a Control-Worker-Judge agent framework with sub-100ms sensor fusion pipelines and leading cross-functional departments.

LLM orchestration Sensor Fusion LLM AI Agent Frameworks Data Pipelines Marketing Automation Robotics

AI Engineer Intern

Microsoft

Feb 2026 - Aug 2026 (6 months)

Selected for Microsoft's AI Innovators Program; architecting an end-to-end AI product using Azure AI, Copilot Studio, and multi-agent frameworks. Participating in weekly system design reviews on LLM integration and agent architecture.

Azure AI CoPilot Studio Multi Agent Frameworks LLM Integrations System Design

AI Engineer Intern, R&D Lead

Callmetric AI

Mar 2026 - May 2026 (2 months)

Fine-tuned large STT models on proprietary call center audio using LoRA, reducing transcription error on domain-specific vocabulary. Pioneered audio-to-visual signal conversion to improve model accuracy on emotionally ambiguous speech.

Stts LoRAs Fine Tuning Mel Spectrograms Model Evaluation

AI Research Intern

Firat University

Jul 2025 - Sep 2025 (2 months)

Designed and built a multi-agent Clinical Decision Support System with 6 specialized agents and a Supervisor-Agent arbitration layer. Deployed a production RAG chatbot over the university's institutional knowledge base.

Multi Agent Systems RAG Clinical Decision Support Chatbot Deployment

Machine Learning Intern

Nexpred Solutions

Jul 2024 - Aug 2024 (1 month)

Delivered production-integrated computer vision models for super-resolution, object detection, and retail loss prevention using YOLOv8, CNNs, and GANs. Built a browser-based hand detection application in TensorFlow.js and React.

Computer Vision YOLOv8 CNN Gan TensorFlow JS React