Natasha Maniar
Conference affiliation: McKinsey & Company · 2025
Related work and projects
Natasha Maniar helped develop MemPal, a wearable memory aid that lets older adults ask where they last left an object. She made the prototype the subject of her MIT master’s thesis and was first author of the 2025 study that tested it in participants’ homes. MIT thesis listing · MemPal paper
On this page
Testing a wearable memory aid at home
Natasha studied computer science and engineering at MIT, graduating in 2023 with a minor in finance. She earned a Master of Engineering in electrical engineering and computer science in 2024. In Pattie Maes’s Fluid Interfaces group at the MIT Media Lab, Natasha worked with postdoctoral mentor Samantha Chan on MemPal and took part in the home trials. Chan, Wazeer Zulfikar, Scott Ren, Christine Xu, and Maes co-authored the paper with her. Ren and Xu helped with user testing and app development. MIT undergraduate commencement program · Undergraduate profile · Media Lab graduate roster · Natasha’s account of the trials
MemPal used images from a camera angled toward the wearer’s hands to build a searchable activity diary. When someone asked where they had left their keys, the system searched that diary and described where it had last seen them. In home trials with 15 older adults, participants searched fewer rooms with spoken guidance than they did unaided, even when the researchers included incorrect responses. With those errors included, the improvement in how often participants found the object fell short of statistical significance. MemPal paper
In her account of the trials, Natasha described a participant who found her medication dispenser’s buzzing unhelpful. MemPal gave her a way to try a reminder tied to something she had just done, such as finishing a meal. For Natasha, involving older adults in the design gave them a say in when a device intended to support independent living should interrupt their routines. Natasha’s account of the trials
Researching how teams divide work with coding agents
In 2025, Natasha worked as a business analyst at McKinsey & Company. She contributed to the survey draft and analysis for its April report on enterprise adoption of open-source AI and collaborated on its November study of how software teams adapt their workflows and roles around AI. April survey report · September interview · November software development study
Natasha and Martin Harrysson used legacy modernization and new feature development to explain two ways of dividing work between people and agents. A modernization project can require an agent to understand much of a codebase while giving it a clearly defined result to produce. People can specify that result up front and review the finished work. When a team is still deciding what a new feature should do, it can ask an agent for several plausible versions and use those prototypes to decide what to build. AI Engineer Code 2025 recording
Read the topics behind these talks
1 conference talk
Loading more talks and highlights…
