▶ Watch ↗AI Engineer World's Fair 202619:53
Conference affiliation: Anthropic · 2026
Malhotra’s August 6, 2022 retrospective describes his Azure authoritative-DNS work from June 2019 through May 2022 and subsequent move to Etsy’s Personalization organization as a staff machine-learning engineer on its Computer Vision team. It discusses slow builds, integration tests, troubleshooting guides and dashboards as opportunities for senior-engineer ownership, and his desire to combine software engineering with machine learning. It also describes research-assistant work at the Information Sciences Institute during his master’s studies; it does not establish the degree’s institution or subject.
bruteforced.devRepository pinned on his reciprocally linked personal GitHub profile, described as the Baby Lizards Problem, a variation on N-Queens. Account hosting is verified; sole code authorship is not.
github.com · Source ↗Pinned repository in his GitHub namespace demonstrating cats-versus-dogs image classification using logistic regression. The retained profile does not establish sole authorship.
github.com · Source ↗Pinned repository described by his GitHub profile as a personal portfolio.
github.com · Source ↗Pinned fork of EagleW/Writing-editing-Network, described as code for Paper Abstract Writing through Editing Mechanism. This establishes possession of a fork, not authorship of the original repository or paper.
github.com · Source ↗Anthropic’s August 18, 2026 article explicitly credits Sachin Malhotra as its author, with contributions from Michael Segner. It describes the team’s Slack-based CI incident responder, repository-managed skills, lessons.md memory and situation reports. Its opening incident concerns roughly 44 missing tests: Claude identified a feature flag, Malhotra asked a colleague to revert it, and Claude verified 3 minutes later that skip rules had been removed and the error rate was back to baseline. The article reports a median 14 minutes to first evidence-grounded incident analysis, describes a 617-line shadow-divergence investigation skill Malhotra built from troubleshooting, and credits the team with ci-weather, which combines incident channels, build metrics, merge-queue statistics and deploy lag into shared reports. These are the author’s reported internal experiences, with team attribution preserved.
claude.comSachin Malhotra builds developer infrastructure and develops practical controls for agents operating production systems. At AI Engineer World’s Fair 2026, he described his work on Anthropic’s CI team: test quarantining, merge automation, CI autoscaling, and merge queues—the machinery engineers use to test, merge, and ship code safely. His approach to agent autonomy replaces broad credential access with operational budgets that limit the volume and pace of actions, account for recoverability, and make mistakes visible.
Malhotra pursued a master’s degree, during which he worked as a research assistant at the Information Sciences Institute, and held engineering roles at Microsoft and Etsy before Anthropic. He joined Microsoft’s Azure authoritative DNS team in June 2019 and left in May 2022 to join Etsy’s Personalization organization as a staff machine-learning engineer on its Computer Vision team. His reflection on his Microsoft years describes his desire to combine software engineering with machine learning. It also emphasizes work that can be overshadowed by feature development: improving slow builds, neglected integration tests, troubleshooting guides, and missing dashboards. He treats recognizing and repairing those weaknesses as part of becoming a senior engineer.
At Anthropic, Malhotra described his team’s Claude-based CI incident responder, which operates through Slack, investigates failures using engineering tools, and produces situation reports. Repository-managed instructions and a running lessons.md file preserve troubleshooting knowledge between incidents. In a feature-flag incident, the workflow connected diagnosis to a human rollback and then checked that the skip rules had disappeared and the error rate had returned to baseline. The agent’s useful contribution included verifying recovery, rather than stopping at an explanation of the failure. His team also built ci-weather to combine incident updates and CI metrics into shared status reports; making those reports readable still required the team’s judgment about communication.
Malhotra’s World’s Fair talk, “Give the Agent a Budget, Not a Token,” explains why useful operational agents need controls beyond prompts and token scopes. He recounts a cleanup operation in which an empty stage in a command pipeline removed a filter, widening the selection to every workload. The agent deleted about 200 workloads in 90 seconds, affecting resources used by about 20 engineers. It was using his credentials and attempting ordinary cleanup. He locates the failure in giving it unbounded power without close supervision: the agent could perform an authorized operation far faster and more broadly than intended.
Malhotra gives written instructions and infrastructure complementary roles. Prompts explain intent and incorporate lessons from incidents; deterministic controls count operations and reject requests that exceed policy. His objective is to let agents own useful operational loops while keeping the power to exceed consequential limits with a human.
▶ Watch ↗AI Engineer World's Fair 202619:53
Loading more talks and highlights…