- BIO
Matt Stancliff
"I've seen things you people wouldn't believe."
Technology Life Highlights
Hosted Dial-Up BBS
1996 - 1999
First Domain Name Purchase
1997
Computer Science BSCS with OS/Networking & Econ & Linguistics Focus Areas
Georgia Tech
2001 - 2004
First OSS Contribution
Contributed small patch to Heartbeat
2006
Helped found Hacker Dojo
Co-signer on building lease, helped with events and buildouts
2009
Improved Web Security
The entire industry was sleeping on what SSL ciphers and DHE params meant for both security and performance. My benchmarks 'went viral' for showing performance differences between cipher suites and DHE parameters and even caused some high profile projects to fix their code for not loading DH params properly historical summary a, historical summary b
2011
Contributor to Redis
Refactored many components for security and performance while building new features focusing on stability and efficiency and community management
2014
Designed & Built a Redis Replacement
After Redis decided to go in a less open and less community focused direction, I spent a couple years building the most memory efficient database possible. End result is the highest performing in-memory database in the world with over 8,000 hours of development on top of a lifetime of working in high performance distributed systems experience, yet nobody really wants to buy it when free worse performing choies exist.
2016-2020
Building Computational Consciousness Actors
Growing non-consensus AI projects with higher ROI than dumping terabytes of data into segmented text writing platforms.
2024
Employment Vibes
Local Computer Repair Shop
Tech / Sales
Atlanta, GA
2000 - 2001
Campus Computer Store
Sales
Atlanta, GA
2001
Regional ISP
Network Engineer / Infrastructure / Platform / Automation / Observability / Data Center Management
Atlanta, GA
2003 - 2005
SMS Aggregator
Infrastructure / Data Center Admin
Atlanta, GA
2005 - 2006
Network Security Company
Linux Net Kernel Module Dev
Atlanta, GA
2006 - 2007
Social Network Company
Infrastructure / Network / Platform / Automation / Observability / Data Center Admin
San Jose, CA
2008 - 2012
Personal Project Travel Break
UK, Spain, Ireland, Costa Rica, US
2012 - 2013
Redis Dev and Sales and Consulting Support
Sales Eng / Dev / Consulting / Support / Community Management
New York, NY
2013 - 2016
Personal Project Travel Break
Vietnam, Bali, NZ, Japan, US
2016 - 2021
Hiring Data Platform
Product Dev / Infrastructure / Network / Security / SOC2 / Admin
Los Angeles, CA
2021 - 2023
Experience
(timeframes are dated from first productive practical experience; inquire for more specifics if curious)
These are notable experience snapshots, but the goal is never individual tool usage, but rather how to combine a wide lifetime of experience across multiple disciplines into building high performance and cost-optimized services, products, teams, and companies (while keeping everything easy to understand and simple to expand over time).
These are the tools and practical experience we use to help build products and services as efficiently as possible.
platform automation
terraform
(3+ years)puppet / facter -> ansible
(14+ years)ssh
(23+ years)telnet / netcat
(27+ years)custom pull/fetch deployment systems
(16+ years)IPN / webhooks
(15+ years)automating distributed TLS ACME cert renewal securely
(7+ years)observability automation
grafana
(3+ years)prometheus / victoria metrics
(3+ years)legacy nagios / cacti / rrdtool / mrtg / custom dashboards
(21+ years)service health check / supervision / smf hierarchies for automatic restarts
(20+ years)time-series analysis, trending, and anomaly detection methods (holt-winters, ARIMA, modern transformers)
(18+ years)Python
python 2.x
(16+ years)python 3.x
(9+ years)async / coroutines / concurrency
(5+ years)dataclasses
(4+ years)fastapi / pydantic
(5+ years)django
(14+ years)pandas
(8+ years)matplotlib
(12+ years)C
C89 / C99
(22+ years)C11
(11+ years)clang-format / clang-tidy
(8+ years)Make
(22+ years)cmake
(15+ years)pthreads
(21+ years)epoll / kqueue / libuv / event-loop concurrency
(14+ years)network APIs / BSD sockets
(21+ years)godbolt
(10+ years)esoteric languages
erlang
(18+ years)scheme / lisp
(23+ years)javascript / ecmascript
(17+ years)css
(25+ years)unicode
(10+ years)smalltalk / squeak
(21+ years)objective c
(19+ years)swift
(9+ years)product development
internal tools platforms
(21+ years)customer-facing application security and usability and design
(22+ years)instrumenting applications with observability hooks
(18+ years)building for long-term platform sustainability and maintainability
(18+ years)productivity
tcsh -> bash -> zsh
(27+ years)vi -> vim -> nvim
(22+ years)screen -> tmux
(21+ years)perl (rip)
(27+ years)git
(16+ years)legacy cvs -> svn -> hg
(21+ years)rsync
(21+ years)services
nginx
(15+ years)apache (2) / mod_* / cgi-bin (rip)
(27+ years)postfix / dovecot
(16+ years)dns hosting (bind / powerdns / djbdns)
(20+ years)hosted dns (dnsmadeeasy / route53 / et al)
(15+ years)modern payment API automation (stripe, etc)
(13+ years)software firewalls and IP networking internals (ipchains / iptables / nftables / pf / crossbar / security groups / ACLs)
(20+ years)virtualization/isolation tradeoffs (VMs vs cgroups vs zones vs chroot vs containers)
(15+ years)databases
(usage and optimization/reliability)
mysql (onprem / self-hosted)
(23+ years)mysql (rds / aurora serverless v2)
(2+ years)postgres
(21+ years)sqlite
(14+ years)clickhouse (self-hosted) / clickhouse cloud
(3+ years)legacy berkelydb (bdb)
(22+ years)custom developed databases
(18+ years)legacy tech business things: RADIUS, Asterisk, voip provisioning automation, proxcard provisioning automation, captive portal automation
(20+ years)scalability / performance
network design and management (subnets / VLANs / VPCs / zones / regions / routing)
(23+ years)server management via iLO / IPMI / serial
(21+ years)distributed processing
(19+ years)AWS basics (S3, EC2, SQS)
(18+ years)AWS intermediates (IAM, billing optimization, redundancy, spot optimizations)
(9+ years)AWS advanced (VPC, bridging VPCs via non-overlapping subnets, ALB vs NLB vs CloudMap vs App Mesh, code build/deploy/artifact, eventbridge, cloudtrail)
(3+ years)AWS automation using boto3
(14+ years)websockets
(4+ years)pubsub architectures
(15+ years)queue architectures / topic subscription modeling
(15+ years)sales engineering
tech authority in sales calls
(10+ years)customize features for custom customers
(9+ years)training other sales and support persons about technical products
(9+ years)Linux
networking / kernel / file systems / architecture / security / scalability
(21+ years)drbd
(15+ years)also various combinations Linux / FreeBSD / OpenBSD / Solaris / macOS
(25+ years)the old ways: init.d, crond, ntpd, supervisord, netstat, ifconfig, iptables
(21+ years)the new ways: systemd, ip, ss, nft
(9+ years)cross-platform multi-host roaming logins via ldap / pam / nsswitch / kerberos / nfs
(19+ years)Design
is the design easy to use? is it easy to understand? does it look nice? is the color scheme effective? does it follow usability principles? does your grid layout work? is interactive latency acceptable?
(21+ years)Modern AI
pytorch; transformers, more than meets the TPU
(2+ years)einops
(2+ years)creating custom decoder-only transformer architectures in pytorch
(2+ years)clean reusable training with LightningCLI
(2+ years)deploying model snapshots into live environments
(2+ years)using openai APIs for GPT assistant output generation
(2+ years)ML
everybody loves a good maximally separating hyperplane
(20+ years)luatorch (rip)
(15+ years)LBP window extractions
(15+ years)opencv
(18+ years)numpy
(17+ years)scikit-learn
(14+ years)general image model usage / detectron2 / etc
(5+ years)futurism
we are constantly living in the future and it keeps getting more future the longer we future
(20+ years)Notable Writings
imagine if somebody did this
Published on: October 30, 2012
just me having clearly unreasonable ideas in 2012 nobody would ever pursue or explore further (record scratch—oh, you mean it's now over 10 years later and implementing these ideas in dozens of companies is now a global $100 trillion industry? go figure.)
a fly on the wall
Published on: July 18, 2014
again more of my clearly unreasonable ideas. look at this weirdo over here thinking 'A.I.' is going to be a big deal. what a waste of time! nobody should listen to this person anymore.
I also talked to multiple VCs in 2014 (thanks to intros from Sam) about funding a GPU-accelerated database platform to grow AI learning capabilities, but all the VCs cared about at the time was 'so does it do blockchain? no? no blockchain? no blockchain no funding.'
what is a feeling?
Published on: October 11, 2011
some historical random musings on what may encode information overlays into our experiences (such as how we can have the same information under different emotional contexts; the information doesn't change, but our 'experience' of the information changes, etc)
Research
Research Summary
Header1
Header 2