There was a time when the mythology of the engineer required solitude: one person, one screen, one heroic quantity of caffeine. Jasmine Oliveira’s version contains rather more people. It begins with a teacher, passes through a laboratory, grows around a crowded lunch table, and eventually arrives at a manager asking engineers what the numbers mean to them. The code matters. So do the conversations around it.
Oliveira grew up in Fall River, Massachusetts, a city whose Portuguese culture she connects to her grandparents’ migration from the Azores. As a freshman in high school, she was learning Visual Basic and doing what curious teenagers do when software first reveals its pliability: messing around, changing things, seeing what would break. Her teacher noticed. He told her that software could be a career and that not many women entered the field. Oliveira took the observation as a challenge.
The challenge became a computer science education at the University of Massachusetts Boston. The surviving artifacts are pleasingly varied. There was EcoSim, a game in which a player tries to keep an ecosystem balanced. There was a wirelessly controlled, four-wheeled catapult programmed in C# and known as the “Marshmallow” Robot. Before engineering became a matter of organizational alignment, it involved the urgent matter of making confectionery airborne.
The sandbox acquires a laboratory
At the Center for Astrophysics | Harvard & Smithsonian, the sandbox became scientific. Oliveira worked on a WorldWide Telescope Android application and contributed to SPECdata, open-source interactive software designed to simplify the analysis of broadband rotational spectra. The problem was not ornamental. Modern instruments could record immense stretches of spectral information, while lines from unknown chemical species hid among familiar compounds. Analysis was slow and laborious. Software could accelerate both recognition and discovery.
In 2017, Oliveira appeared with Marie-Aline Martin-Drumel and Michael McCarthy on the program of the International Symposium on Molecular Spectroscopy. Their subject was SPECdata. The work foreshadowed much of what came later: take a dense field of signals, make it intelligible, and help a human decide what deserves attention.
A career in four translations
After university, she also worked on a machine-learning-powered marketing conversion platform. Then, in April 2019, a former colleague, Eli Daniel, told her about a young company named Jellyfish. Oliveira met the team over lunch. There were roughly ten people. They showed her the product, and she liked both. An early startup often asks a candidate to bet on an unfinished thing. Oliveira took the bet.
She joined as a full-stack engineer and called herself a generalist. The designation was literal, but not merely technical. Alongside building software, she occupied a role Jellyfish called “Dev Custodian.” She fielded requests from across the company, sorted priorities, watched the state of projects, and kept product and customer activity in view. Her concise description was better than any job architecture: making sure balls did not get dropped.
The company outgrew the table
The early Jellyfish office had a large communal table. Lunch there felt, Oliveira recalled, like Thanksgiving dinner: everyone together, with banter and interesting conversation. The company grew. Functions appeared that had not existed when she arrived, including sales and marketing. The table became an organization, and the organization produced a familiar modern difficulty. Its engineering work was expensive, essential, and surprisingly hard to describe.
Oliveira’s career expanded with that difficulty. She led teams building tools that helped organizations understand engineering resource allocation and streamline software capitalization, and she became a software engineering manager. Her current title, founding engineer, brings the arc neatly back toward building, although the work around the work has never disappeared.
The last number requires explanation. Asked in 2021 to describe her team culture in three words, Oliveira offered four: “supportive, thoughtful, silly, transparent.” The miscount was deliberate enough to keep. It also says more about her managerial vocabulary than a tidy trio would have. Precision has its uses. People sometimes require an extra word.
Remote work tested that culture. Jellyfish already had a Thursday ritual in which each person said something they were grateful for. Oliveira’s team added a counterweight: an “airing of grievances.” Complaints could concern a broken development environment or the misery of wet socks. The ceremony gave irritation a harmless stage. A team cannot spend its whole life in dashboards. Occasionally it needs to complain about the rain.
The trouble with counting puddles
Rain supplies Oliveira’s sharpest analogy for engineering measurement. Counting issues, pull requests, or story points, she wrote, is like measuring rainfall by counting puddles. It may offer a rough idea, but it misses rather a lot. A flurry of activity can conceal unplanned work, maintenance, customer support, blocked projects, or a strategic change. The convenient number is not always the consequential one.
Metrics also have politics. For an individual contributor, measurement can feel less like illumination than surveillance. In her 2023 LeadDev West Coast talk, Oliveira addressed that fear directly. Managers, she argued, could use development metrics in collaboration with engineers, creating more transparent conversations about job satisfaction, self-advocacy, burnout prevention, and team dynamics. The crucial preposition is with.
Her later writing put the approach into practice. A Jellyfish developer-experience initiative combined survey signals with operational information and prompted investment in test automation. The published case study reported that satisfaction with test automation doubled; engineers were satisfied with nine of ten topics measured; overall satisfaction and well-being rose 15 percent. The important move was not admiring a score. It was changing the environment behind it.
That habit now extends into artificial intelligence. Oliveira has worked on Jellyfish’s AI-impact initiatives and has publicly framed the task as moving beyond the flood of commentary toward real data and real stories about how coding tools alter engineering. In October 2024, she moderated a Boston fireside chat with QuotientAI co-founder Julia Neagu on engineering management in the age of AI. The new tools changed; the old question endured. What signal can leaders trust?
Serious work, unserious hiding places
Oliveira’s public persona is too playful for the severe priesthood of metrics. She co-hosted Jellyfish’s Off the VPN podcast with Allison Regna, turning workplace dilemmas into brisk debates: the engineer who automates a job, the senior colleague with opinions on everything, the harmless collaboration that someone decides looks suspicious. The format allows technical culture to admit that it is also office culture, which means motives are mixed, etiquette is consequential, and absurdity is never far away.
Her favorite early-office story confirms it. When a change of snack vendor produced a Pop-Tart shortage, employees began hiding the remaining packets like treasure. One went into the cold-brew keg. Oliveira put hers in the freezer. It is a small anecdote, but useful profiles are built from small anecdotes. A person who can discuss software capitalization and hidden pastries in the same career has escaped the curse of becoming a corporate abstraction.
Outside work, she has described herself as an avid hobbyist. She reads, plays video games, goes to the gym, and loves Kizomba enough to have flown to social dances and classes. Her musical list has included Lauryn Hill, Giveon, and Joé Dwèt Filé. In college she volunteered with Strong Women, Strong Girls, mentoring elementary-school girls and helping coordinate site leaders. Since 2023, her LinkedIn record lists her as a mentor with Hack.Diversity. The teacher’s high-school remark about women in software did not end as a private challenge. It became something she could help other people answer.
What the dashboard cannot hold
Oliveira’s path is not a departure from engineering into “people work,” as if people were an optional plugin. It is a widening definition of the system. A product has users. A team has dependencies. A company has incentives, anxieties, stories, rituals, and occasionally wet socks. Ignoring these variables does not make an engineering leader more rigorous. It merely leaves the model incomplete.
This is why the scientific beginning matters. Spectral analysis asks the observer to separate meaningful lines from a complicated field. Engineering intelligence asks much the same. Oliveira has spent her career moving between those acts of translation: instrument to researcher, code to customer, activity to business outcome, metric to human conversation. The job is not to erase complexity. It is to make complexity navigable.
The ten-person lunch table will not return. Growth rarely permits such charming geometry. But its virtues - directness, humor, trust, a willingness to hear the whole table - can survive if somebody builds for them. Oliveira’s work suggests that measurement, used carefully, can be part of that construction. Count what matters. Keep the context. Ask another question. And perhaps check the freezer before declaring the Pop-Tarts gone.