|
1. Who is August Forsman?
1.1 Introduction
I'm an AI Engineer living in Stockholm with broad technical expertise spanning from microcontrollers to supercomputer clusters. I thrive working in the intersection of AI, data and software engineering.
1.2 Education
- • Engineering Physics, Uppsala universitet (class of 2018)
- • Rymdgymnasiet, Space Engineering and Freestyle Skiing (class of 2011)
- • Uppsala Musikklasser, Song (class of 2005)
1.3 What I've worked with
LLM, agents, NLP, information retrieval, data pipelines, statistical analysis, computational science and high-performance computing.
1.4 Technical skills
Python, Pydantic AI, C++, C, SQL, FastAPI, Docker, Azure, BigQuery, OpenWebUI, HuggingFace, Git, Rust, Terraform, Linux (Arch, btw)
1.4 Technical skills
Python, Pydantic AI, C++, C, SQL, FastAPI, Docker, Azure, BigQuery, OpenWebUI, HuggingFace, Git, Rust, Terraform, Linux (Arch, btw)
1.5 FEM 2D Medical Torus
Figure 2: Diffusion! Cool!
A simulation study of a torus-shaped medical implant in the course Applied Finite Element Methods at Uppsala universitet.
The reaction-diffusion is modeled through a system of PDEs which is discretized and solved using the Crank-Nicolson method.
Figure 2: Diffusion! Cool!
1.5 FEM 2D Medical Torus
A simulation study of a torus-shaped medical implant in the course Applied Finite Element Methods at Uppsala universitet.
The reaction-diffusion is modeled through a system of PDEs which is discretized and solved using the Crank-Nicolson method.
1.6 2D Ising Model
Figure 4: Magnetization per spin
A Markov chain Monte Carlo simulation study of a ferromagnetic material as part of the course Computational Physics at Uppsala universitet.
The comparison of the Metropolis-Hastings and Heat-bath algorithm is computed using a C++ templatized implementation of the NumPy library.
Figure 4: Magnetization per spin
1.6 2D Ising Model
A Markov chain Monte Carlo simulation study of a ferromagnetic material as part of the course Computational Physics at Uppsala universitet.
The comparison of the Metropolis-Hastings and Heat-bath algorithm is computed using a C++ templatized implementation of the NumPy library.
1.7 MPI Performance Study
Figure 3: Increasing CPUs at fixed load
Part of BSc thesis. A comparison of one- and two-sided MPI for Conway's Game of Life with toroidal boundary conditions.
The boundaries are computed by exchanging the halo region between the nearest neighbours. Experimental data is obtained by running the program on the UPPMAX supercomputer cluster.
Figure 3: Increasing CPUs at fixed load
1.7 MPI Performance Study
Part of BSc thesis. A comparison of one- and two-sided MPI for Conway's Game of Life with toroidal boundary conditions.
The boundaries are computed by exchanging the halo region between the nearest neighbours. Experimental data is obtained by running the program on the UPPMAX supercomputer cluster.