Bojan Jakimovski
Machine Learning Engineering Lead at Loka
Bojan Jakimovski contributes to Loka's engineering publication as a Machine Learning Engineering Lead. Their published work covers ESM-C, AWS Trainium, Protein Language Models, Benchmarking.
Published work
Hardware benchmarkScoring 42 Million Protein Variants a Day on AWS Trainium2
João Correia, Telmo Felgueira, Tiago Gonçalves, Bojan Jakimovski, Jim Burtoft, Louise Ping · August 18, 2026
A native PyTorch benchmark showing EvolutionaryScale ESMC-300M running on AWS Trainium2: 490.8 variants per second on a single trn2.3xlarge — approximately 42.4 million protein variants scored per day at approximately $1.26 per million, 60% below H100 in this fixed-shape D2Deep benchmark.
Benchmarking Grok 4.3 on Amazon Bedrock Mantle vs. the xAI API: The Engineering Walkthrough with Loka
Nina Cvetkovska, Bojan Jakimovski · July 10, 2026
The main post covers what we found: Grok 4.3 lands in the same 96–97% accuracy band on GSM8K whether you reach it through Amazon Bedrock Mantle or the first-party xAI API, and low reasoning effort captures most of the…
Provider benchmarkThe Many Paths to Grok 4.3
Nina Cvetkovska, Bojan Jakimovski · July 10, 2026
A public-facing benchmark story on Grok 4.3 through Amazon Bedrock Mantle and the xAI API, focused on accuracy, latency, cost, reliability, and enterprise deployment tradeoffs.
Benchmarking GPT-5.5 on Amazon Bedrock vs. the OpenAI API: The Engineering Walkthrough with Loka
Petar Kalinovski, Bojan Jakimovski · June 15, 2026
The main post covers what we found: GPT-5.5 on Amazon Bedrock matches OpenAI API accuracy and comes out faster on every latency and throughput dimension we measured. This post covers how we ran it. If you want to…
Model evaluationHow Loka Evaluates and Builds with Frontier Models on AWS
Petar Kalinovski, Bojan Jakimovski · June 15, 2026
A comparison of GPT-5.5 through Amazon Bedrock and the OpenAI API across answer quality, latency, throughput, and reliability, followed by two production-shaped copilot builds.
Accelerator benchmarkRunning Hugging Face Carbon on AWS Trainium2
Bojan Jakimovski, Loka Applied Research · May 21, 2026
A practical benchmark showing Hugging Face Bio Carbon running on AWS Trainium2 with NxD Inference, covering what worked, what we measured, and why it matters for bio and HCLS teams moving open models into production.
Running Hugging Face’s Carbon on AWS Trainium2 with NxD Inference
Bojan Jakimovski · May 21, 2026
Hugging Face Bio’s Carbon release is interesting for two reasons at once. First, it is a biology model. Second, it is not an infrastructure outlier. Many genomic models come with custom architectures, specialized…
The Agent Assembly Line
Mario Petkoski, Bojan Jakimovski, Zafir Stojanovski · March 2, 2026
Productionizing Agentic Use Cases in Weeks, Not Months
Deploying Trinity-Mini-DrugProt-Think on Amazon SageMaker AI
Bojan Jakimovski, Petar Kalinovski · February 23, 2026
If you work in regulated domains (Healthcare, Life Sciences, Finance) you routinely hit constraints that break the default “just call a hosted API” approach:
Reinforcement learningPost-Training an Open MoE Model to Extract Drug-Protein Relations
Bojan Jakimovski, Petar Kalinovski · February 23, 2026
An ablation-driven study of GRPO-style reinforcement learning with LoRA on Trinity Mini for biomedical relation extraction, covering the training choices that materially changed performance.
Part 2: Agentic Patterns 101 with Loka & Strands-Agents
Nina Cvetkovska, Petar Kalinovski, Bojan Jakimovski · September 2, 2025
After exploring the foundations of Agentic Patterns in Part 1, where we looked at more structured workflows like Sequential and Parallel workflows, LLM Routing, and Reflection, this time we are going to push further…
Part 1: Agentic Patterns 101 with Loka & Strands-Agents
Nina Cvetkovska, Petar Kalinovski, Bojan Jakimovski · August 27, 2025
These days, Agents are everywhere in daily conversation and it seems everyone wants one. We read or hear about AI Agents, Agentic AI, Agentic Architectures, and Agentic Patterns almost every day. But what do these terms…
Part 4: Benchmarking DeepSeek Cost, Performance and Business Use Cases on AWS
Crhistian Cardona, Bojan Jakimovski · April 10, 2025
A DeepSeek benchmark on AWS that compares cost, performance, and deployment choices for business workloads using open-weight models.
Part 2: Deploying Distiled DeepSeek-R1 Models on Amazon SageMaker AI
Crhistian Cardona, Bojan Jakimovski · February 20, 2025
A deployment walkthrough for serving distilled DeepSeek-R1 models with Amazon SageMaker AI, including the AWS setup for open-weight reasoning models.
Part 3: Deploying DeepSeek-R1 Models on AWS- designed Silicon Instances
Crhistian Cardona, Bojan Jakimovski · February 20, 2025
Engineering notes on deploying DeepSeek-R1 models to AWS-designed silicon instances and serving open-weight reasoning models on AWS.
Part 1: Deploying Distilled DeepSeek-R1 Models on Amazon Bedrock
Crhistian Cardona, Bojan Jakimovski · February 17, 2025
Deploy distilled DeepSeek-R1 open-weight models through Amazon Bedrock, with a hands-on AWS setup from the Loka engineering team.
Dive into DeepSeek
Crhistian Cardona, Bojan Jakimovski · February 12, 2025
How the open-recipe LLM is transforming GenAI