AI Developer tech skillsets.
Darko Medin
Data Scientist and a Biostatistician. Developer of ML/AI models. Researcher in the fields of Biology and Clinical Research. Helping companies with Digital products, Artificial intelligence, Machine Learning.
While these skills may vary according to the role, i will discuss the most significant ones that almost every AI developer needs to have. There are so many discussions today what skillset are related to actual AI Developer Experts, so i decided to make a summary which may be expanded and its easier for anyone interested into AI skillsets to understand them better. This article is written by Darko Medin, the creator of bioaiworks.com (a Bio AI platform)
1. MODEL DEPLYMENT. I put this as number 1 technical skill for AI EXPERTS, because its the one that turns an AI models into a valuable asset which can solve problems... This involves deploying models locally, testing them, deploying on servers, working with serves and cloud a lot. To many, may be surprising, a lot of work in the domain of model deployment is done using command prompts. There are automated platforms, where a point and click is used, however, technical skills in creating server instances, virtual machines, installing software, running models on servers using bash or cmd are a must in my opinion.
2. 'CONTAINERIZATION OF SOFTWARE' . Most AI models will need to be packaged into some form of software, wither client side or server side and container software such as Docker or Kubernets is almost a must.
2. Creating NEURAL NETWORKS. Most AI models today, including Transformers, LLMs, Reasoning models, Outcome predicting AI, Recommender systems, Image recognition, Voice Recognition, almost all present day are mostly based on Deep Learning/Reinforcement using Neural Networks. Enough said about the importance about being able to create Neural Networks.
3. PYTHON PROGRAMMING. Working with Data, Creating Neural Neural Networks, Training them, Evaluating and significant part of Model deployment usually is mostly done using Python programming language. Most AI implementations today rely on Python libraries such as PyTorch, TensorFlow, Keras, sklearn and Transformers, that brings me to the next one. Its best optimized for this, can implement 'under the hood' other programming languages such as C++ and Rust for speed and is a general purpose programming language for AI. Python became the most popular programming language today on GitHub, mostly driven by AI development and knowing Python is an essential AI skill today.
4. STATISTICS/DATA SCIENCE. Creating AI models is a lot like Statistical modeling and involves a lot of Data Science. Not discussed much, but this actually makes Statisticians and Data Science able to program, ideal candidates for developing AI models if augmented by other skills needed. Not the mention the fact that AI model evaluation and cross-validation is a typical Data Science / Statistical skill and that Neural Networks are inherently Statistical Learning Algorithms (just very complex ones).
5. SOFTWARE ENGINEERING. While most fear that AI will take software engineering jobs, to develop AI we need a lot of Software engineering. Id say AI will only increase the demand for software engineering who understand how AI works and can develop AI models in collaboration with Data Scientists and AI engineers.
6. WEB DEVELOPMENT. Believe it or not, one of the most important skills for AI experts is the realm of Web development. Almost all AI products involve working with Frontend, Backend, Servers, Cyber security web dev and so on.
7. PROBLEM SOLVING. Problem solving is almost everywhere in AI projects. Most AI projects revolve around a specific problem AI is supposed to resolve and you as a professional need to understand how. Then almost every phase of AI model development and deployment.
8. DOMAIN KNOWELDGE Without domain knowledge, its different to problem solve and manage AI projects. Domain knowledge is a must for any AI project.
9. DATA ENGINEERING. Last but not least. AI projects are at least 50% working with Data, Databases, Data lakes and data pipelines. This tells for itself why Data engineering is an important AI skill.
10. SPECIAL SKILLS - Cognitive functions, Neuroscience, Reasoning in General, Causal reasoning, abstraction principles, AI agents, GitHub of course! PS using AI to Augment you as an AI developer. Don't underestimate this one. Finally, one of the most important skills? Understanding GPUs, TPUs, parallelization of data processes. There is even more, Cybersecurity to some extent.
by Darko Medin
Creator of BioAIWorks
Onco data —> AI. ??