<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Nicolas Essipova on Medium]]></title>
        <description><![CDATA[Stories by Nicolas Essipova on Medium]]></description>
        <link>https://medium.com/@essipova?source=rss-a2910a1a83de------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*BejF5WqlDj-mLXSrEeEo4w.jpeg</url>
            <title>Stories by Nicolas Essipova on Medium</title>
            <link>https://medium.com/@essipova?source=rss-a2910a1a83de------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 10:20:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@essipova/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Roles in Data Science: The Definitive Guide]]></title>
            <link>https://medium.datadriveninvestor.com/roles-in-data-science-the-definitive-guide-2fea0a2be421?source=rss-a2910a1a83de------2</link>
            <guid isPermaLink="false">https://medium.com/p/2fea0a2be421</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[management]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Nicolas Essipova]]></dc:creator>
            <pubDate>Sun, 14 Jun 2020 23:25:36 GMT</pubDate>
            <atom:updated>2020-06-20T23:28:15.523Z</atom:updated>
            <content:encoded><![CDATA[<h4><strong>Do you wish for a complete guide detailing what the different roles in data science do, and how they differ from one another? Here it is!</strong></h4><p>Several decades ago, it was common for someone who works with programming to simply refer to themselves as a programmer, whether they worked with developing websites, desktop software, embedded systems, or a vast array of other programming applications. These days, we have specific terminologies to convey what type of programming we do, whether it is web developer, site reliability engineer, native Android or iOS developer, et cetera.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1LQt3fUOBd24__D0iQeMzA.jpeg" /><figcaption>Image by <a href="https://unsplash.com/@sctgrhm">Scott Graham</a> on <a href="https://unsplash.com/">Unsplash</a>.</figcaption></figure><p>However, due to the advent of rapid industrialization of artificial intelligence and big data, many professionals (both technical and non-technical alike) are scrambling to make sense of all the hype and how to best describe what it is we data professionals do. For a while, the title of data scientist has been used as an umbrella term similar to how the title of a programmer used to be. In this article, I aim to make life easier for as many as possible, by clarifying exactly what the different titles imply and common titles that they may often be mixed in or confused with — which I will refer to as “overlap”.</p><p>As a disclaimer, please note that it is more common than not to hold numerous roles in a single job position, which is a significant contributing factor to the confusion around roles, and we sometimes hear stuff like:<br>“Data scientists are experts in AI” (<a href="http://businessoverbroadway.com/2018/02/18/a-majority-of-data-scientists-lack-competency-in-advanced-machine-learning-areas-and-techniques/">which is not true</a>) and that “machine learning engineers are full-stack data scientists who can do data engineering” (also not true, which we will get into).</p><p>Complete list of roles covered in this article — feel free to skip ahead to the ones most relevant to you or your organization:</p><ul><li>Data Analyst</li><li>Data Scientist</li><li>Decision Scientist</li><li>Actuarial Scientist</li><li>Data Engineer</li><li>Data Architect</li><li>Machine Learning Engineer</li><li>Machine Learning Scientist</li><li>Machine Learning Researcher</li><li>Machine Learning Architect</li><li>Data Science Architect</li><li>Business Intelligence Analyst</li><li>Quantitative Analyst</li><li>Quantitative Developer</li><li>Artificial Intelligence Specialist</li></ul><h3>Data Analyst</h3><p>A great place to start defining a data analyst would be to see what differentiates an ordinary analyst from an ordinary scientist, as the same principle applies. Data analysts primarily curate insights from data by looking at what is known, usually to answer specific business questions that have been put forth. They generally do not work with predictions or generalization, considering “what if” scenarios, pattern recognition, or conducting experiments. Their use of mathematics, programming, and machine learning are limited to the basics.</p><p>Overlap:<br>Junior Data Scientist.</p><h3>Data Scientist</h3><p>The most succinct and apt description of the data scientist role is as follows:<br>“Being a data scientist requires a hybrid of statistics, programming, and business sense. It is a real scientist who can explore insights from the data for a company’s business needs.”</p><p>There are a lot of parallel with the data analyst role, though data scientists have a greater emphasis on conducting experiments, predicting future outcomes, generalize existing data and new data, recognize patterns, utilize mathematics to find deeper insight, train and optimize machine learning models, and able to collect, clean, and process data with basic data engineering. Good visualization and presentation skills are a big plus.</p><p>Let us be clear though: they are not and do not have to be experts in artificial intelligence. In fact, <a href="http://businessoverbroadway.com/2018/02/18/a-majority-of-data-scientists-lack-competency-in-advanced-machine-learning-areas-and-techniques/">a majority of data scientists lack competency in advanced machine learning areas and techniques</a>. Anecdotally, most of my students who have gone on to work as data scientists for Fortune 500 companies had very limited knowledge of machine learning, because they really did not need more than the basics.</p><p>Machine learning is a tool, and data scientists just need to know when to use them, how to use them, and when to try alternative options — Much like how a carpenter does not need to understand the intricacies of how his chainsaw works or how to build one from scratch. It might be helpful, but not necessary.</p><p>Overlap:<br>Data Analyst, Decision Scientist, Machine Learning Engineer, Junior Data Engineer.</p><h3>Decision Scientist</h3><p>Interestingly, decision scientists are what most businesses would benefit most from having in their organization — and yet they are a rare breed.</p><p>Data scientists excel at applying programming and mathematics to solve business problems, but they typically lack strong business acumen, ability to understand and define business problems, know how to effectively communicate with different stakeholders within an organization, and do not have an in-depth understanding of decision-making processes within organizations. However, decision scientists tick all of these boxes and more.</p><p>While data scientists provide the framework for machines to make the correct decisions, decision scientists provide the framework for humans to make the right decisions. However, this does not mean that they do not possess expertise in mathematics, programming, and technology — because they certainly do. Decision scientists are qualified to work as data scientists. Simply, their focus is more on humans and less on machines.</p><p>If you’d like to know more, I recommend following <a href="https://medium.com/@kozyrkov">Cassie Kozyrkov on Medium</a>. She’s currently the Chief Decision Scientist at Google.</p><p>Overlap:<br>Data Scientist, Data Analyst, BI Analyst</p><h3>Actuarial Scientist</h3><p>Actuarial scientists are sometimes mentioned within the big data domain, though they follow a very different path towards their first employment and focus on very specific problems. In most countries, you must demonstrate competence by passing a series of rigorous professional examinations.</p><p>They possess expertise in mathematics and programming, which they apply to model matters of uncertainty. However, this is where their similarities end with data scientists and quantitative analysts — as their primary focus is on dealing with evaluating risks and maintaining the economic stability of insurance or financial organizations. To do so, they have a deep understanding of economics, finance, investment, and insurance. They can work at almost any company where financial risks must be evaluated, and are often found at banks, investment firms, hospitals, and manufacturing.</p><p>Overlap:<br>Data Scientist, Quantitative Analyst.</p><h3>Data Engineer</h3><p>Quite simply, data engineers are responsible for digestion, management, and access to data within an organization — along with testing and maintaining data architectures. Sometimes they are responsible for designing the entire architecture as well. They are absolutely vital to any organization that wishes to make use of their data for analytics purposes.</p><p>Data engineering is an impressively diverse domain, as data engineers can further specialize in dealing with data streaming (Kafka, Faust, Spark Streaming), cloud as opposed to on-prem (AWS, GCP, Azure, DigitalOcean), and machine learning deployment (KubeFlow, Feature Stores, MLOps)</p><p>Unfortunately, it is common for organizations to delegate these responsibilities to data scientists. Most data scientists are woefully undertrained relative to data engineers to undertake data engineering tasks beyond basic ETL (Extract, Transform, Load). If you do not have a data engineer: get one before you consider getting a data scientist.</p><p>Overlap:<br>Data Architect, Machine Learning Engineer, DevOps Engineer, Quantitative Developer.</p><h3>Data Architect</h3><p>Data architects are specialized in designing and implementing database systems that fulfill the specific needs of their organization. They work in tandem with data engineers in building the data infrastructure, wherein the data architect creates the blueprint while the data engineer executes it, very similar to the relationship between architects and engineers in the housing and construction industry. However, it must be noted that data engineers often do the work of a data architect if their organization does not have access to one, and many data architects are experienced, data engineers.</p><p>Typically to be recognized as a data architect, it is strongly advised to demonstrate competence by passing a professional examination such as by <a href="https://aws.amazon.com/certification/certified-solutions-architect-associate/">Amazon Web Services</a>, <a href="https://cloud.google.com/certification/cloud-architect">Google Cloud Platform</a>, or <a href="https://docs.microsoft.com/en-us/learn/certifications/azure-solutions-architect">Microsoft Azure</a>.</p><p>Overlap:<br>Data Engineer, Machine Learning Engineer, DevOps Engineer</p><h3>Machine Learning Engineer</h3><p>Plain and simply, machine learning engineers are primarily concerned with the deployment, maintenance, and versioning of machine learning models for production purposes. They share many of the same responsibilities as a <a href="https://en.wikipedia.org/wiki/Site_Reliability_Engineering">Site Reliability Engineer</a>, except they work within the machine learning domain. Being a machine learning engineer does not necessarily mean developing nor optimizing machine learning models, as that is usually either done by the data scientist or machine learning scientist. However, it is not uncommon for a machine learning engineer to also work as a machine learning scientist.</p><p>Machine learning engineers work predominantly with cloud services such as AWS SageMaker, GCP AI Platform, and Azure Machine Learning, and are expected to possess basic competence in data engineering and good conceptual knowledge of machine learning.</p><p>Overlap:<br>Data Scientist, Machine Learning Scientist, Quantitative Developer, Data Engineer.</p><h3>Machine Learning Scientist</h3><p>Machine learning scientists are concerned with designing, building, and optimizing machine learning systems. They are expected to have an intimate understanding of mathematics, programming, and machine learning. They regularly hold a master’s degree or doctoral degree in machine learning or another quantitative field and spend a significant amount of their time on reading relevant research papers and experimenting with new ideas, to ensure they are always providing the best solution to their organization.</p><p>As machine learning is a very wide and deep domain, it is common for machine learning scientists to specialize in specific subdomains, such as computer vision, natural language processing, time series forecasting, parallel computation, statistical modeling, and general deep learning.</p><p>Machine learning scientists that are specialized in deep learning are aptly titled deep learning engineers, or deep learning scientists. Unfortunately, it is estimated by <a href="https://www.elementai.com/news/2019/2019-global-ai-talent-report">Element AI</a> that there are fewer than 30 000* deep learning experts worldwide, and the supply problem is difficult to fix. There is a serious lack of approachable education and most online courses barely scratch the surface of this domain, thus a significant amount of further reading of academic papers is required, which are not approachable to most enthusiasts. However, this does not seem to stop many from self-proclaiming themselves to be deep learning experts.</p><p>*: The real figure is much smaller than this. The data used by Element AI include self-proclaimed experts that have only taken introductory courses in university and online courses, such as on Coursera. As stated, these just scratch the surface of the domain, and therefore these individuals are not considered experts by respectable institutes and organizations. Thank you for prompting me to take a closer look at the data, Geoff Hulten.</p><p>Overlap:<br>Data Scientist, Machine Learning Engineer, Deep Learning Engineer, Deep Learning Scientist, Machine Learning Researcher, Quantitative Analyst.</p><h3>Machine Learning Researcher</h3><p>What differentiates a machine learning scientist from a machine learning researcher is that the latter works with pushing the human knowledge further and beyond what is already known. While machine learning scientists are concerned with building what is best for their organization, a machine learning researcher is only concerned with discovering new knowledge — regardless of whether it has any business application or not.</p><p>Typically it only makes sense for an organization to employ machine learning researchers if they are able to allocate massive amounts of money to accelerate their AI R&amp;D without causing any economic instability and there is no urgency in producing business applicable results, which is why you mostly see machine learning researchers working at larger institutes such as Google, Facebook, Microsoft, and Nvidia.</p><p>Machine learning researchers almost always hold a doctoral degree in machine learning, though it is not impossible to get into this position with a master’s degree in machine learning if the individual has a few years of industry experience working with responsibilities that fall under the role of a machine learning scientist and good publication record.</p><p>Overlap:<br>Machine learning scientist, deep learning engineer.</p><h3>Machine Learning Architect</h3><p>Machine learning architects are primarily concerned with designing and managing the entire machine learning project, with emphasis on what is best for their organization in regards to costs, performance, and reliability.</p><p>They are typically machine learning engineers and/or scientists with several years of experience, and they understand the full intricacies on how to set up and optimize a machine learning system end-to-end, what possible problems and mistakes may arise and how to solve them, and know how to take their solution from a prototype to fully deployed into production with a <a href="https://en.wikipedia.org/wiki/MLOps">supporting lifecycle system</a>. In addition, they also possess great customer empathy to ensure they can identify the pain points.</p><p>If you’d like to learn more about working as a machine learning architect, I recommend <a href="http://blog.intelligentsystem.io/2020/04/thinking-like-a-machine-learning-architect/">this article by Geoff Hulten</a>, a senior manager from Google.</p><p>Overlap:<br>Machine learning engineer, machine learning scientist, data architect.</p><h3>Data Science Architect</h3><p>Similar to machine learning architects but differ in a similar manner as data scientists differ from machine learning scientists.</p><p>Data science architects are concerned with orchestrating a data science project from start to finish, which involves designing, implementing, and improving end-to-end data science solutions. They understand the business problem and how to set up the appropriate analytics workflow to find the insights needed for their organization. Responsibilities include identifying what data to collect and how to process it, which statistical analyses to consider, summarising results as a decision with a confidence band, and communicate the decision(s) supported by the project to stakeholders.</p><p>Unlike machine learning architects, there is less emphasis on aspects regarding machine learning and instead greater emphasis on statistics and finding the insights that satisfy the criteria to solve their business problem.</p><p>Information regarding the data science architect role was defined and formulated as a result of a discussion with <a href="https://www.linkedin.com/in/santona-tuli/">Santona Tuli, Ph.D</a>. She is currently a research data scientist and team lead at CERN. Credit goes to her.</p><p>Overlap:<br>Data scientist, machine learning scientist, machine learning architect, machine learning engineer, decision scientist.</p><h3>Business Intelligence Analyst</h3><p>Business intelligence analysts, or just BI analysts, are concerned with identifying market and business trends so that their companies may increase profitability and efficiency, and they do so by analyzing data retrospectively. They are a lot more similar to data analysts and often join organizations as consultants to do descriptive analysis and standard reporting. They are focused on KPI (key performance indicators), trends, and business reports.</p><p>Overlap:<br>Data analyst</p><h3>Quantitative Analyst</h3><p>Quantitative analysts, or “quants”, apply advanced mathematics and statistics to solve financial and risk management problems. They work with developing complex models for their firms so that they can make risky financial and business decisions about problems such as investment and pricing. How they differ from actuarial scientists, in particular, is that quantitative analysts are a lot more concerned with predicting outcomes using any relevant data that they can find, like data scientists, whereas actuarial scientists work more with historical and stable data to draw conclusions.</p><p>Traditionally, quantitative analysts have utilized stochastic calculus, differential equations, and probability-based models to solve their problems — but it is becoming increasingly more popular to employ talented data scientists and machine learning scientists for this role, due to the massive amounts of data and compute power that financial institutions have access to, and the promising reliability of machine learning in producing results.</p><p>Overlap:<br>Data scientist, machine learning scientist, actuarial scientist, quantitative developer.</p><h3>Quantitative Developer</h3><p>Quantitative developers serve a similar role to both data engineers and machine learning engineers. Their main concerns are to develop and maintain the necessary infrastructure required by institutions to enable quantitative analysts to do, well, quantitative analysis.</p><p>However, it is not uncommon for job descriptions to describe quantitative developers as analysts, and there is often a significant amount of overlap between quantitative analysts and developers.</p><p>Overlap:<br>Data engineer, data architect, machine learning engineer, quantitative analyst.</p><h3>Artificial Intelligence Specialist</h3><p>Honestly, this can mean anything. I’ve seen an individual with a bachelor’s degree in business who have taken some courses on Udemy and Coursera proclaiming themselves as an AI specialist, and I’ve seen people with PhDs in mathematics and genuinely impressive knowledge of machine learning calling themselves AI specialists as well.</p><p>Personally, I would use this title for SEO and communicative purposes, as I do not expect most people to know what machine learning scientists or engineers do — hence the purpose of this article. If you see someone holding this title: ask them to elaborate on their competencies.</p><p>Overlap:<br>Almost anything.</p><h3>Final Remarks</h3><p>Prior to publishing this article, I made sure to ask some of my LinkedIn connections to verify that their roles were accurately described. In addition, this article was inspired by <a href="https://www.youtube.com/watch?v=gDg-tUACuvQ">this informative video by Geoff Hulten</a>, who currently works as a senior manager at Google and machine learning scientists. This article aims to complement his video. He also provided a lot of valuable feedback on this article, for which I am grateful!</p><p>If you find anything incorrect, I’d be very glad to receive your feedback. If you have any questions or simply wish to connect, feel free to find me on LinkedIn. I’m easy to find.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2fea0a2be421" width="1" height="1" alt=""><hr><p><a href="https://medium.datadriveninvestor.com/roles-in-data-science-the-definitive-guide-2fea0a2be421">Roles in Data Science: The Definitive Guide</a> was originally published in <a href="https://medium.datadriveninvestor.com/">DataDrivenInvestor</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>