Senior Software Engineer · Cadence

Moisey Alaev

I ship AI that people can rely on.

I'm a senior software engineer at Cadence, working where AI engineering meets backend engineering: I design the agents and LLM systems, then build the services that keep them dependable. Before that I put reinforcement learning agents in charge of cooling AI data centers at Phaidra.

Email me GitHub LinkedIn Resume

The orange line under “rely” is the agent's latest temperature trace. Watch it calm down as the agent learns.

Fig. 1 · training live in your browser

Episode0

Agent in band–

Thermostat–

learning agentplain thermostatyousafe band

A small agent is learning, from zero, to keep a simulated server rack in its safe temperature band while spending as little cooling as it can. This is a toy (tabular Q-learning, about 60 lines). At Phaidra I shipped the real version: a PyTorch RL agent for liquid cooling in AI data centers.

CurrentlyCadence

RoleSenior Software Engineer

MaterialGo · Python · PyTorch

BasedCentral Valley, CA

In progressM.S. CS, Georgia Tech

FocusLLM systems · RL · ML infra

Experience5 years · 3 startups

Sheet 1 of 2Off the clock →

95%LLM routing accuracy on thousands of weekly healthcare tasks

−20%GPU temperature spikes with an RL cooling agent

−60%issues escalated to engineering, via Slack AI agent skills

days → 1 hrcustomer onboarding validation, with an ML commissioning CLI

#1most-used internal dashboard at Phaidra, built and owned

Work

Five years of putting ML behind software people can depend on.

Fig. 220212022202320242025202620272028

Cadence

Phaidra

Secureframe

HADD research

UCLA, B.S.

Georgia Tech, M.S.

Hover or tap a row ·current rolepast rolesdegreedegree in progress

95%routing accuracy across thousands of weekly tasks

Cadence

SENIOR SOFTWARE ENGINEER · APR 2026 – NOW

Rebuilt the LLM task router for a healthcare company: a distributed Go service that reads free-text requests, applies customer context and assigns the right follow-up workflow. Every model change passes labeled evals, offline replay and a shadow deployment before it ships.

~1,500 interactions automated weeklyup to 20 hours of operations work saved weekly−60% escalations to engineeringissues resolved ~1 week sooner

GoLLM evalsAWS EKS · SQSagent skillsSnowflake · Datadog

−20%GPU temperature spikes

Phaidra

SOFTWARE ENGINEER → ENGINEER II · NOV 2023 – APR 2026

Productionized a PyTorch reinforcement learning agent for liquid cooling in AI data centers. Promoted to Engineer II for leading work across two ML teams and the Connectivity team, and was technical lead for the simulation platform used to test cooling scenarios before deployment.

onboarding validation 2–4 days → 1 hour#1 most-used dashboard+15% ML training data quality−50% manual testing time

PyTorch RLFastAPI · ReactGo · gRPCPostgreSQL

+10NPS points, from device-management automation I drove

Secureframe

SOFTWARE ENGINEER · JUL 2022 – NOV 2023

Automated third-party integrations for security compliance and built the Slack and Jira backend behind in-product messaging and ticketing.

+15% sales wins over competitors300 bugs and infra issues resolved−40% customer complaints

RailsGraphQLTypeScript

~5%experimental error

HADD ML Research Group

SOFTWARE CO-LEAD · JAN 2021 – SEP 2022

Co-led the team behind an open-source Python module for semi-supervised tensor decomposition, tested on raw data such as EEG video.

PyTorchTensorFlowNumPy

Peer review

What the people I've worked with say.

Reviewer 1 · Phaidra One of the most impactful ML engineers I have worked with during my time at Phaidra.

Giuseppe PintoPhysical AI Engineer at Phaidra, Ph.D.
Senior colleague · October 2026

Read the full review

Moisey is one of the most impactful ML engineers I have worked with during my time at Phaidra.

At Phaidra I led a small team to work on a physical AI project, and I asked him to join as the technical lead. From day one he owned the architecture and drove the end to end implementation. Furthermore, he was on the forefront of agentic workflows, cutting our development cycle significantly.

What sets Moisey apart is how he thinks about products, he naturally identified the decisions that made the product shine without over-engineering. The best parts of working with him were his proactivity and his willingness to dive into unknown terrain to make an impact.

I would recommend Moisey without hesitation to any team looking for a ML engineer who combines deep technical ownership with real product instinct.

Reviewer 2 · Research An exceptional researcher who works well in teams, is an excellent leader.

Jamie HaddockAssistant Professor of Mathematics, Harvey Mudd College
Research advisor · October 2021

I give my highest recommendation to Moisey. Moisey has been my research student for about eight months and was a student in my Machine Learning (mathematics) course in Fall 2020. He is an exceptional researcher who works well in teams, is an excellent leader (often he directs a group including several graduate students and faculty years his senior), and has picked up new skills and highly challenging concepts for our research project quickly. Moisey led our team in reimplementing our Python library from a PyTorch backend to a Tensorflow backend. This was a challenging task as it required rethinking some of the subroutines used due to varying levels of sparse data support across these two backends. He additionally taught himself how to work with challenging multi-modal video data using openCV and produced many useful graphical visualizations of the data for analysis. The work that he has completed has produced a manuscript that will be submitted soon.

Reviewer 3 · Secureframe Patient and helpful without an ego.

Hayley Schluter LarrabeeStaff Software Engineer
Teammate · July 2023

Moisey had been at Secureframe for almost a year when I joined, and he was really helpful ramping me up not only to the company, but also in introducing me to Ruby which I had never worked with before.

Working on large features together, we would often pair program, with Moisey helping fill me in on my gaps in backend knowledge. We'd leave the sessions with a working solution, which was great! I always felt comfortable sharing with him when I didn't understand something well, and he's patient and helpful without an ego. He's also proactive with writing unit tests, a big help for the overall maintainability of our code.

Moisey is a very friendly and collaborative teammate who's easy to work with, and a really down to earth person who isn't phased by the shifting priorities that come with agile development. I'd love to work with him again!

All recommendations on LinkedIn ↗

Moisey Alaev

Five years of putting ML behind software people can depend on.

Cadence

Phaidra

Secureframe

HADD ML Research Group

Why I sit between software engineering and machine learning.

Evals are the test suite.

Data rots quietly.

Agents earn their autonomy.

What the people I've worked with say.

Latest on github.com/moiseyalaev

Still a student, on purpose.

Georgia Tech

UCLA