How to Address Misconceptions in Data Roles

The 5 biggest myths about getting your first data role 👇 Everyone talks about how to break in. But most advice? It’s outdated or just plain wrong. 𝐋𝐞𝐭’𝐬 𝐛𝐮𝐬𝐭 𝐬𝐨𝐦𝐞 𝐦𝐲𝐭𝐡𝐬: 1. 𝐘𝐨𝐮 𝐧𝐞𝐞𝐝 𝐭𝐨 𝐤𝐧𝐨𝐰 𝐞𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠 - No, you need to show you can learn quickly. - Curiosity > Perfection. 2. 𝐘𝐨𝐮 𝐧𝐞𝐞𝐝 𝐚 𝐦𝐚𝐬𝐭𝐞𝐫’𝐬 𝐝𝐞𝐠𝐫𝐞𝐞 - Not always. - Projects, proof of work, and storytelling matter way more than fancy degrees. 3. 𝐘𝐨𝐮 𝐦𝐮𝐬𝐭 𝐛𝐞 𝐠𝐫𝐞𝐚𝐭 𝐚𝐭 𝐋𝐞𝐞𝐭𝐜𝐨𝐝𝐞 - Data roles ≠ Software Engineering. - Focus on SQL, data cleaning, EDA, and business impact. 4. 𝐘𝐨𝐮 𝐧𝐞𝐞𝐝 𝐲𝐞𝐚𝐫𝐬 𝐨𝐟 𝐞𝐱𝐩𝐞𝐫𝐢𝐞𝐧𝐜𝐞 - Internships, freelance work, open-source, or even side projects count. - Start where you are. 5. 𝐀𝐩𝐩𝐥𝐲𝐢𝐧𝐠 𝐨𝐧𝐥𝐢𝐧𝐞 𝐢𝐬 𝐞𝐧𝐨𝐮𝐠𝐡 - Nope. The best opportunities come from conversations, not just applications. - Referrals can 10x your chances. 𝐁𝐫𝐞𝐚𝐤𝐢𝐧𝐠 𝐢𝐧 𝐢𝐬 𝐭𝐨𝐮𝐠𝐡, 𝐛𝐮𝐭 𝐧𝐨𝐭 𝐢𝐦𝐩𝐨𝐬𝐬𝐢𝐛𝐥𝐞. Stay consistent. Build real projects. You’re closer than you think. 💬 𝐖𝐡𝐢𝐜𝐡 𝐦𝐲𝐭𝐡 𝐝𝐢𝐝 𝐲𝐨𝐮 𝐛𝐞𝐥𝐢𝐞𝐯𝐞 𝐚𝐭 𝐟𝐢𝐫𝐬𝐭? ♻️ Save it for later or share it with someone who might find it helpful! 𝐏.𝐒. I share job search tips and insights on data analytics & data science in my free newsletter. Join 12,000+ readers here → https://lnkd.in/dUfe4Ac6

During a recent mock interview with someone preparing for a data engineer role, we ran into a super common challenge job seekers face: The job description mentioned machine learning and stats, but my mentee only had basic experience in those areas. And with a big presentation coming up, we had to figure out the best way to prepare without wasting time or getting overwhelmed. Here’s what we did: 1) Focus where it matters most I told them: Imagine it’s a test. 90% of the questions are on things you know well, 10% on stuff you don’t. Where would you spend your time? Obviously, on the 90%, that’s where you can really shine. 2) Play to your strengths Instead of cramming complex ML topics overnight, we doubled down on what they already knew: cleaning data, building solid pipelines, writing scalable code—all stuff that directly applies to the job. Then we built a strong presentation around that. 3) Address the gaps without pretending We practiced how to talk about the ML/stats side in a smart way: not by faking expertise, but by showing how their data engineering skills help data scientists succeed. It’s not about having every skill; it’s about showing how you add value to the team. 4) Be real, but show you're a problem solver We also talked about how to confidently say: “I don’t know X yet, but here’s how I’d figure it out or approach it with the tools I do have.” This shows honesty and the mindset every team wants: someone who takes ownership and knows how to solve problems. The result? They walked into that interview feeling confident and clear on how to present themselves. Key reminder: You don’t need to know everything. You just need to show how what you do know makes you a great fit. So here's a question for you: If you have ever faced a situation where you were underqualified in one area, tell us in the comments how you used your strengths to stand out. ➕ Follow Jaret André for daily data job search tips 🔔 Hit the bell icon so you never miss a post

𝐀 𝐂𝐨𝐦𝐦𝐨𝐧 𝐌𝐢𝐬𝐜𝐨𝐧𝐜𝐞𝐩𝐭𝐢𝐨𝐧 𝐀𝐛𝐨𝐮𝐭 𝐃𝐚𝐭𝐚 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 🚨🚨 𝐃𝐚𝐭𝐚 𝐀𝐧𝐚𝐥𝐲𝐭𝐢𝐜𝐬 is a rapidly evolving field crucial in decision-making across various industries. However, several misconceptions can hinder its understanding and effective use. When I first came across the data analytics field, I thought it was all about collecting and analyzing data, which made me overlook the critical importance of context, interpretation, and the human element in the analytics process. After diving deeper into it, I understood that while data collection and analysis are fundamental components, they are only part of a larger picture that includes: 📌 understanding the business problem, 📌 defining clear objectives, and 📌 effectively communicating findings. The following were noted throughout my learning journey: 𝐓𝐡𝐞 𝐈𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐜𝐞 𝐨𝐟 𝐂𝐨𝐧𝐭𝐞𝐱𝐭 📌 The context in which data is collected and analyzed is vital for deriving meaningful insights. 📌 Without a clear understanding of the business objectives or the specific questions that need to be answered, data analytics can lead to misleading conclusions. 📌 Analysts must be able to interpret data within the framework of the business environment, industry trends, and stakeholder needs. 𝐓𝐡𝐞 𝐑𝐨𝐥𝐞 𝐨𝐟 𝐈𝐧𝐭𝐞𝐫𝐩𝐫𝐞𝐭𝐚𝐭𝐢𝐨𝐧 📌 Data analytics requires critical thinking and domain knowledge to translate data into actionable insights. 📌 Analysts must be skilled in storytelling with data, presenting findings in a way that resonates with decision-makers and drives strategic actions. 𝐓𝐡𝐞 𝐇𝐮𝐦𝐚𝐧 𝐄𝐥𝐞𝐦𝐞𝐧𝐭 📌 I realized that the human element in data analytics cannot be underestimated. 📌 Collaboration among teams, communication of insights, and the ability to adapt to changing circumstances are essential for successful data-driven decision-making. 📌 Relying solely on automated tools and algorithms can lead to a disconnect between data insights and real-world applications.   𝐎𝐯𝐞𝐫𝐚𝐥𝐥 📌 Effective data analytics requires a comprehensive understanding of context, strong interpretative skills, and a collaborative approach. 📌 By recognizing these elements, organizations can harness the full potential of data analytics to drive informed decision-making and achieve their goals. What is another misconception you have heard or had about data analytics? Please share it in the comments section. #EdwigeSongong #ESAnalysis #DataAnalytics #DataStorytelling

AI agents are disrupting data engineering and we're wasting time arguing if it's a fad, instead I'll tell you why you're already halfway to be an AI engineer. I've had 100+ conversations in the last 8 months with companies who struggled to build AI agents. One common theme emerged - they chose to ignore their data engineers and use all the new and cool tools on the market to build and deploy these agents. We’re all getting sucked into the vortex of the AI hype machine and as with any new tech, a slew of new tools emerge promising ease of use and fast results 🤩 But they're completely unnecessary since data engineers already built the infrastructure and systems to support AI agents 🥳 There are two types of agents: 1. AI agents - apps that use LLMs to make decisions or perform tasks 2. Code (non-AI) agents - apps you code to make decisions or perform tasks Agent chains is another new idea for connecting multiple agents together in a workflow to complete a multi-step or complex task. From a data engineering standpoint, agents are like Airflow tasks and agent chains are like Airflow DAGs. No need for fancy “agentic” tools, you got what you need already. Another common misconception is that you need to build a RAG system to support AI agents. Not really… RAG is a fancy term for look-up. Your data platform is already set up to provide AI agents a way to look up information it needs when constructing prompts. #Iceberg works with Python, Rust and Go. Snowflake, Databricks, Redshift and other query engines provide APIs to execute SQL queries programmatically. You can start by collecting metadata about your data assets, which may already be in your catalog, and use it for similarity searches and enrichment when creating prompts. You don’t need a vector database or to create embeddings. Do this to help your company build AI agents in a mature way: - Make your data accessible via APIs - Build code wrappers around common data processing tasks - Make it easy to deploy tasks on your Airflow - Provide backend (DB) store to enable agent memory Personally, I’m excited about the possibilities of AI agents. I believe using the tools we already have, we can quickly bring reliability, scale and performance today, without waiting for the fancy new AI tools to mature. p.s. as a DE, what about AI agents confuses you the most?