AIMultipleAIMultiple
No results found.
Alparslan Polat

Alparslan Polat

AI Researcher
4 Articles
Stay up-to-date on B2B Tech

Alparslan Polat is an AI Researcher at AIMultiple. His current work focuses on GPUs, synthetic data generation, computer vision, and large language models (LLMs).

Professional Experience

During his long-term internship at the Turkish Petroleum Corporation (TPAO), Alparslan actively participated in a computer vision project, making significant contributions to the generation of synthetic data.

Research Interest

At AIMultiple, Alparslan contributes to projects involving GPUs, synthetic data generation, computer vision, fine-tuning, and large language models (LLMs). He is eager to explore the reasoning capabilities of LLMs, develop benchmarks for testing and inference, and investigate their potential applications in various fields, including geological research.

Education

  • Master’s Degree in Data Informatics (Starting September 2025), Middle East Technical University

  • Bachelor’s Degree in Geological Engineering, Middle East Technical University

  • High School Diploma in Mechatronics, ASO Technical College

Latest Articles from Alparslan

AISep 8

GPU Concurrency Benchmark

We benchmarked the latest NVIDIA GPUs, including the NVIDIA (H100, H200, and B200) and AMD (MI300X), for concurrency scaling analysis. Using the vLLM framework with the gpt-oss-20b model, we tested how these GPUs handle concurrent requests, from 1 to 1024.

AISep 3

Top Image Recognition Tools Compared

We evaluated the real-world performance of top cloud vision tools for object detection tasks by benchmarking their default API configurations across 5 classes. This included contrasting performances, analyzing features, and comparing service offerings in relation to pricing. Benchmark Results Performance overview at IoU=0.

AIAug 4

LLM Pricing: Top 15+ Providers Compared

We analyzed 15+ LLMs and their pricing and performance. LLM API pricing can be complex and depends on your preferred usage. If you plan to use: Hover over model names to see their full names and over headers to see explanations about the columns.

DataSep 16

Synthetic Data Generation Benchmark & Best Practices

We benchmarked 7 publicly available synthetic data generators sourced from 4 distinct providers, utilizing a holdout dataset comprising 70,000 samples, with 4 numerical and 7 categorical features, to evaluate their performance in replicating real-world data characteristics. Below, you can see the benchmark results where we statistically compare the synthetic data generators.