This as-told-to essay is based on a conversation with Abhinav Bohra, a senior applied scientist at Amazon. He’s in his 30s and lives in Seattle. The following has been edited for length and clarity.
For most of my career, everything I built lived and died inside Amazon.
I joined Amazon in 2019 as a data scientist and eventually became a senior applied scientist in 2024. My background is in AI recommendation systems, and I work on models that help determine which products shoppers see on Amazon.
While I was building a reputation inside the company, I wasn’t building much of one outside it. People in the industry outside Amazon were publishing research papers on the same problems I was solving, and I never joined the conversation.
I eventually realized that without a public record of my work, I was missing opportunities to speak at industry events, serve on conference committees, and build a reputation that could help me if I ever needed to find a new job. When it came to professional opportunities outside Amazon, it felt like I was starting from zero.
The lesson I learned was that internal reputation doesn’t travel. My biggest career regret is how long I took to show up outside the company. I’d be a few years further along if I’d realized this sooner.
I developed rules for myself on how I could share my work outside Amazon
At first, I was hesitant to publish externally. I treated anything within a mile of my day job as radioactive, which meant I shared nothing publicly. There are processes for publishing company work externally, but they can require documentation and review, and I didn’t want to go through that process.
Eventually, I realized I could do external work without using the company’s proprietary data or information. The data was proprietary, but my ideas and research were not.
Take a problem I worked on a few years ago: How do you train a good AI model when you have almost no labeled data? Research teams everywhere hit that wall.
I developed some rules for myself. If a problem exists in published research, it’s a conversation. If it only exists in our codebase, it’s not. Nonpublic data, internal metrics, customers, code, roadmaps, and unreleased products are all off the table. If I’m still unsure, I won’t share it.
I published outside research and boosted my LinkedIn presence
It surprised me how much meaningful technical discussion was still possible without revealing anything proprietary.
In 2021, I started taking steps toward sharing more of what I was learning, and I published my first research paper around a year later. Publishing papers opened the door for me to start reviewing other researchers’ work and helped me build relationships in the industry.
Eventually, I was able to land some external speaking opportunities. I think having published research and experience reviewing other researchers’ work is what made the difference.
Posting on LinkedIn became part of my strategy, too. I shared my research, technical ideas, and conference work. I wasn’t trying to become an influencer — I was just trying to build a professional identity that wasn’t tied to a single employer.
More opportunities have come up since I started sharing my work
In recent years, the going-public bet has started to compound in ways I can actually point to rather than just describe.
I’ve since become a Challenge Co-Chair for ACM RecSys 2026, an international conference focused on recommender systems. I reached out, explaining my background, research, and reviewing record, and said I was interested in serving on the organizing committee.
After the organizers reviewed my background and we spoke on a call, I was appointed. Reaching out got me considered, but the work behind the initial email is what made the opportunity possible.
I’ve also joined a mentorship program through Johns Hopkins, where I advise early-stage ventures on AI and technical strategy.
The outside work has made my inside work better
Building an external reputation isn’t only about having accomplishments to put on my résumé.
I think the work I was doing inside Amazon would have gotten better faster if I’d started engaging with people outside the company earlier.
When you explain your thinking in a paper or on a conference stage, you’re writing for people with no context and no reason to be polite. The weak spots in your work show up fast. Strangers make the best reviewers precisely because they owe you nothing.
Over the years, my outside work has made me more confident at Amazon. I think the biggest change was becoming more comfortable contributing an idea before I had every detail perfectly figured out.
Keeping up with AI and professional development can feel like a second job
AI learning can become an unofficial second job. It’s valuable and exciting, but it also blurs the boundary between professional development and personal time. Outside my day job, I spend evenings and weekends researching retrieval and recommendation systems and submitting research papers on the topic.
I enjoy the work, so I don’t want to frame it as purely negative. But it does affect work-life balance. There are evenings or weekends when I should probably be disconnected, but I end up spending time on AI because I know the field is moving quickly.
AI creates opportunity, but it also creates a learning tax. For people working close to the field, continuous learning has quietly become part of the job, even when it happens outside the job.
Tech workers should build something outsiders can see
There’s value in having a public record of your work if your career ever takes an unexpected turn. Life can be uncertain. There could be layoffs. There could be a difficult job market that makes it hard to find a new job.
You don’t want to spend years developing tools, solutions, and expertise, leave a company, and feel like you’re starting from scratch because everything you built was proprietary and you can’t share it.
My advice to other people working in tech is to make sure you have a profile outside your company that allows someone to understand the value you can bring.
That doesn’t mean sharing proprietary information or violating an NDA. There are other ways to build a public record. For me, that means publishing independent research, speaking at industry conferences, reviewing other researchers’ work, and becoming involved with organizations related to my field.
I wish I’d started sooner.
Do you have a story to share? Reach out to the reporter via email at jzinkula@businessinsider.com, or via Signal at jzinkula.29.
Read the full article here
