Essential AI Tools for Engineers and Content Creators
I wanted to talk about some of my favorite AI tools that I use on a daily basis as a software engineer and a content creator. The three main areas that I’ve started leveraging AI to improve my workflow are coding, research, and writing. I’ll talk about a few AI tools for each of those categories and also give you some honorable mentions that I don’t necessarily use every day but the ones that have been useful whenever I’ve needed them.
My name is Utav. I’m a software engineer based in Seattle. I have about 20 years of experience in the industry where I’ve held diverse software engineering roles and created a few tech startups and I’m currently at Microsoft. My goal is to help you get the best out of your career by mentoring you around five key pillars.
Career Mentorship and Professional Pillars
These pillars are technical skills, engineering efficiency, mindset, entrepreneurship, and Financial Freedom. For coding it’s the usual suspect GitHub Copilot. GitHub Copilot is an AI powered code assistant developed by GitHub in collaboration with OpenAI. It uses machine learning specifically OpenAI’s Codex model to help developers write code more efficiently. My favorite thing about Copilot is that I don’t need to do anything extra to actively use it. It’s in there in the background learning the code as I’m writing it and automatically pops up with useful suggestions at the right moment.
Practical Coding with GitHub Co-pilot
I do a lot more architecture design these days and much less of coding so I forget the syntaxes and semantics of some simple things like disposing a disposable object. Copilot will automatically detect that and suggest that I add a function to dispose it. If I’m writing a helper function and I or someone else has already written it somewhere else in the code Copilot would detect that and suggest that I use that instead. It’s nothing life-changing as in Copilot isn’t going to be threatening my job anytime soon but it definitely helps when you want to code as fast as possible to churn out a quick proof of concept or an MVP.
Safety Precautions and Tabnine
Even though it can help greatly increase your coding speed I recommend that you double check the code that you’ve added via Copilot or any other AI assistant before you release it to production. It is an AI model after all and lacks the contextual intelligence that us humans have so it can occasionally have bugs and security vulnerabilities. Another useful tool is Tabnine which is quite similar to Copilot. They have been around for longer than Copilot and were one of the very first movers in AI coding assistance space. I used to use Tabnine before I switched over to Copilot and it was pretty good. They offer the same key features as Copilot contextual code completion support for a vast array of programming language and works with your favorite IDE.
Privacy Options and Local Inference
One key difference between Copilot and Tabnine is that Tabnine offers more flexibility in terms of privacy with their local inference models for those who prefer to keep their code private. Tabnine definitely leans towards helping you complete the code that you’ve already started as opposed to Copilot which can generate code from scratch. I don’t think one is significantly better than the other it just comes down to preference. Some folks have also suggested Kite as an option to me but I’ve never gotten around to trying it so it wouldn’t be fair to include it here without having had hands-on experience. Feel free to try it if that one tickles your fancy. While not directly related to coding I found Brave to be one of the better browsers for code search.
Brave CodeLLM and Programming Queries
It uses CodeLLM a tool designed for programming related queries that combines comprehensive search results with summarization and explanatory capabilities of large language models. CodeLLM provides AI generated code snippets based on search results complete with step-by-step explanations and source citations for reference and validation. This combination presents search results in a format similar to official documentation which adds the benefit of citations to deepen your understanding of the topic. As a software engineer I frequently have to read academic papers and as a content creator I also have to research about topics that I’m making videos about. During the initial phase of the research process I face two challenges when it comes to using AI tools.
The Challenges of AI Research
First is to identify which AI tool to use because there are so many of them around these days for chat for image for video you name it. The second since different assistants are based off different models their answers can vary so I need to compare them against one another to be able to collect accurate data points to dive deeper into. For that reason for my initial research I use Poe which solves both these issues. It’s essentially one AI app to rule them all. Poe essentially lets you interact with all of the best AI products in one place. With Poe you can engage with the leading large language models like Claude 3.5 Sonnet GPT-4.0 Gemini 1.5 Pro and creating images with state-of-the-art image generators.
Creating Custom Bots and Multi-Bot Chat
These generators include DALL-E 3 Ideogram and Stable Diffusion 3. You can also create your own unique AI powered bots on Poe reaching an audience of millions and generating revenue through their creator monetization program. For my research I really like the multi-bot chat which lets me compare answers from multiple bots in one conversational thread. I find this very powerful because I can not only understand the strengths and weaknesses of different models but also discover and combine information from multiple different models at the same time. I can also leverage the strengths of various bots for specific tasks like ask Llama 3 Groq to summarize articles.
Advanced Research with Specialized Models
I can upload a complex graph to Claude and use it to help me understand the graph or upload a 100 page PDF to Gemini 1.5 Flash and ask it questions about the contents of that file. All of this for the same cost as one ChatGPT subscription. Once I have a decent start on my initial research and I want to dive deeper I use Elicit. Elicit is an AI research assistant that can help you find summarize and organize academic papers and research data more efficiently.
Evidence-Based Decisions and Literature Review
While it is primarily designed for researchers data scientists and academics to streamline the literature review process I have found it quite useful for diving deeper into topics from my initial research helping me make more evidence-based decisions by synthesizing insights from multiple papers and identifying key information. It has been super useful in streamlining my process of finding and understanding academic research. The last thing I do on a regular basis is write whether it’s script for these videos or things like emails and documentation. For code documentation I use Mintlify which is an AI power tool that automates the creation of high-quality context aware code documentation.
Writing and Documentation Automation
It integrates with all of the popular IDEs supports multiple programming languages and generates real-time customizable documentation directly from codebases. Mintlify has been super useful in helping me save time ensuring that my documentation is consistent and of high quality which is really beneficial for onboarding and code maintenance. Most of my scripting work for videos I make is done within Notion which has built-in AI assistant to improve writing. This helps me write a draft very quickly then use the improve writing option to clean up and streamline the content. While this works quite well for bits and pieces the tokens are limited and more heavy usage would require a subscription which is quite expensive.
Notion AI and Apple Intelligence Integration
These days most applications that allow you to write already have some form of built-in AI tool to assist you or you can use something like Poe and use one of the writing bots to help you with your writing. Since I do most of my work on a Mac I’m looking forward to when Apple Intelligence will soon roll out which will have writing assistant built in. Since it is within the ecosystem I believe that it will have a much better integration and perhaps the best AI solution for writing within a Mac. I also want to provide some honorable mentions for AI driven tools that I’ve used before and had a good experience.
Honorable Mentions for Software Engineering
These are tools that aren’t directly related to coding and not something that I would use every day. First up is DeepCode an AI based tool that helps identify bug security vulnerabilities and performance issues in your code. It uses machine learning to analyze your code in real time and provides intelligent suggestions to improve code quality and security. DeepCode supports multiple programming languages and integrates with all popular IDEs help you write cleaner more secure and optimized code. Next up is Harness.io which is a continuous delivery as a service platform that leverages AI and machine learning to automate and optimize the software delivery process.
Optimizing Software Delivery with Harness.io
It helps teams deploy applications faster and more reliably by automating tasks such as deployment testing and roll backs while also providing real-time monitoring and cost management. Harness also integrates with most existing CI/CD pipelines and cloud environments making it easier for teams to deliver high-quality software with minimal manual intervention.