Senior data science manager at US healthcare company Optum, Gilson has worked with data for more than a decade, specializing in data engineering and analytics. Njengengxenye yomsebenzi wakhe ungumholi weqembu, wengamela abantu abasebenza ezinseleleni eziningi.

“Leading a team working on complex problems means balancing people, priorities and technology every day. I often start the morning by looking at what needs attention and what the team needs most.”

People management is just one skill that Gilson uses throughout his role as chief data science officer, as part of a broader talent acquisition strategy.

Imagine reading hundreds of pages of a contract to find one clause, comparing it to two hundred versions and deciding which version is the most similar. Manje phinda lokho ngezigaba ezingama-300, nsuku zonke. Ngokusobala, kunendlela engcono yokwenza, ikakhulukazi uma onjiniyela bedatha nososayensi bedatha behilelekile. Yileyo nselelo engaqala ngayo cishe eminyakeni emithathu edlule. Kusukela lapho, ngisebenzise amathuluzi okucubungula amadokhumenti ukuze ngiguqule izinkontileka zibe amadokhumenti anesakhiwo esincane. But identifying the correct clause among large volumes of text requires advanced techniques: text processing, embedding, expression, segmentation and synthesis. Machine Learning helps with ambiguity detection, noise reduction and clause labeling, while artificial intelligence deals with text fragmentation, making all machine learning tasks possible.

Ngisho namakhono obuchwepheshe athuthuke kakhulu adala kuphela umthelela wangempela lapho ehlanganiswa ne-ecosystem ebanzi. Ngenkathi ukucubungula nokuhlaziya amadokhumenti kubalulekile, isiyalo esivumelanisa zonke izingxenye siwubunjiniyela. Ukucabanga kokuhlaziya nokuklama kuqinisekisa ukuthi izixazululo zisebenza ngaphandle komthungo, ngokulinganisa nokuqina emnyombweni wazo. They guide the selection of the best method and tools for each task, improving performance while reducing processing overhead and uncertainty. At Optum, we value collaboration, empathy and communication, because building innovative healthcare solutions takes more than just code.

One of the hardest parts of working in artificial intelligence and machine learning is understanding their limitations and knowing where they should and shouldn’t be used. I-AI ingaxazulula izinselelo eziningi eziyinkimbinkimbi, kodwa akuyona inhlamvu yesiliva. Kwethula ukungaqiniseki enqubweni nasemiphumeleni, ikakhulukazi uma okulindelekile kungenangqondo. Ukufunda ngomshini, nakuba kubikezelwa futhi kwamukelwa kabanzi, akugonyiwe emaphutheni. Noma kukhethwa imodeli engcono kakhulu, ukushunwa kwamapharamitha, nokuhlanzwa kwedatha okuqinile, imiphumela engamanga isavela.

Enye inselele iza lapho kuhlanganiswa ubuhlakani bokwenziwa nokufunda komshini kusendleleni. Each layer introduces its own uncertainties and when the output of one becomes the input of another, achieving reliable and repeatable results becomes more difficult. Ukuphatha lobu bunzima, ngenkathi kuqinisekisa ukuqina nokuthembeka, kungenye yezici ezinzima kakhulu zensimu. My way of managing this crisis is to stay in the moment, read research papers, explore new methods and engaging technology platforms. I absolutely love a good challenge. Kuyakhuthaza ukuba sendaweni eyazisa ilukuluku futhi esekela abantu abafuna ukweqa imingcele futhi baqhubeke nokufunda.

Sonke sinemikhuba yethu yokugxila, mhlawumbe ikhofi eliqinile nama-headphone akhansela umsindo ukuze sibhekane nama-imeyili eviki. Kodwa uma umsebenzi udinga ubuhlakani nokunemba, ukugxila kukodwa akwanele. We need inspiration. Kimina, lokho kuvela ekushisekeleni: hhayi ukujaha ukuphelela, kodwa ukulwela ikhwalithi nokusimama. Passion fuels curiosity, drives me to explore solutions, support others and keep ideas flowing, creating an impact that elevates teams and transforms businesses.