Customers taught me the job first. My first role, at Dish Network's headquarters in 2013, kept me close to the contact centers, listening to customers and the agents who talked with them all day. In 2016 I joined Sling TV, the first subscription live-TV streaming service in the U.S., in its second year. I held five roles in eight years, including Head of Customer Experience, where we cut the cost of care 78% while satisfaction went up. Then Tubi (Fox; 110M monthly active users), where I owned product operations for internal tools. Twelve years in TV with a decade in streaming, and it all comes back to the same question: where is this breaking, and who has to feel it before someone fixes it?
The receipts
Eight things I do
Open any one to see the role, the year and what happened.
01
Product Operations
Customer and quality data turned into 270+ production fixes, KPIs and SLOs set with product managers, reporting leadership can act on. Unglamorous. Load-bearing.
TubiSr. Product Ops Manager
Covered the PM gap, and named it
Acted as PM for the creator program in lieu of a dedicated PM, and flagged the missing role in writing as a staffing gap that tooling couldn't close, so it wouldn't be inherited silently. I carried the half-year plan overview and the weekly product and cross-functional syncs.
Sling TVEng Leader, Product Ops & SRE
Customer data, into the sprint
Built the product-ops practice for engineering at Sling. CX and quality-of-experience data surfaced the likely bugs; we confirmed them, sized scope and risk, and got them prioritized against the live sprint schedule. 270+ production fixes shipped that way.
Sling TVSr. Manager, Care Tools
A team of PMs, and the A/B program
Managed the team of product managers over the care suite: the chatbot, the help site, and agent tools for chat and voice. I owned the A/B testing program too. Help-bot flows, website designs, call-routing patterns.
Sling TVManager II, Product Ops & NOC
Where product ops started for me
Established the Product Operations team at Sling — the people who sat with product managers to set KPIs and SLOs, then tracked them.
Sling TVManager, UAT
A team of four, from scratch
Created Sling's first user-based testing team. The scripts I inherited tested what the PRD said a feature did. Not what a customer would actually do with it. A new suite written around user behavior caught thousands of issues before release. The team tested cart, offer and billing releases too.
TubiSr. Product Ops Manager
Put H1 on paper, and tracked the creator work
For H1 planning I turned a 1:1 doc into the top projects, each with goals and stakeholders, and wrote the program descriptions for Creatorverse, Jira Admin and TV-pod support. On the creator side, I tracked Creator Portal through its upload-process ship and filed the Creator Hub epics on mobile and web. Two trackers, too. I maintained the Creator Viewer Experience Tracker and built out the Creator Workstream Tracker in Notion.
The last thing I built at Tubi
The handoff. A START HERE guide, the state of each workstream with its locked decisions, an action register with owners, the architecture and engine docs, and a de-personalized version of my own operating rules.
Sling TVEng Leader, Product Ops & SRE
Which SLAs had teeth
Directed SLA work across the engineering teams, with the violation process that ran when one was missed. The signals came out of New Relic, Kibana, and PagerDuty. Incident SLAs were enforced. Performance and reliability SLAs were advisory.
Sling TV
Live, ads and DVR
While DVR had no PM, product ops picked up the work. I prioritized the live and ads bug stream short, medium and long term.
02
AI & Automation
I ran the team that rebuilt the chatbot in front of a contact center back when that took real work. The instinct hasn't changed. Build the agent, then build the thing that stops it shipping on its own — the guardrail in code, the audit that catches the paraphrase that dropped a safety rule, the memory that asks before it writes.
TubiSr. Product Ops Manager
Skills and scheduled tasks for recurring work
My AI work at Tubi sat inside product ops. I used it to streamline retros, labor-cap tracking, migrations and leadership reporting, and enabled others from there. I built ten Claude Code skills, plus four Claude Desktop scheduled tasks: the incident recap, P0/P1 metrics, the weekly update draft and the creator scan.
Enforced the no-auto-ship rule in code
A logged hook gated Jira, Slack, Notion, Drive, GitHub, mutating SQL, scheduled work and local deletes. The first version did the intent judgment itself and grew to 1,364 lines. Wrong architecture. I cut it to 453 by moving the intent judgment to the model's classifier.
Sling TVSr. Manager, Care Tools
Containment past 90%
Inherited Pypestream, Sling's customer-facing chatbot. My team rebuilt it from the ground up while the vendor was still iterating on the product underneath. We worked the most common customer workflows hard, and containment passed 90%.
TubiSr. Product Ops Manager
Ten agents, auditing my own skills
The audit found one guardrail paraphrased about 30 different ways, and one copy of it missing the destructive SQL verbs. I centralized every copy.
Other teams adopted these
PRD, tech-review and experiment-spec skills I built on Claude. PMs and engineers across the tech org picked them up.
Wrote a style guide for updates
An early AI smoke test produced invented impact language. A reporting style guide came out of that, and weekly logs and leadership updates were written to it.
Built it, then kept it observe-only
The creator scan read 8+ creator Slack channels every day and matched what it found against the feature list in Notion. The match quality was unproven. So it wrote nothing anywhere.
Managed an AI vendor end to end
The vendor was Resolve AI. I connected it to Tubi's internal infrastructure so feedback shaped its tools.
Took the AI work past my own desk
Generalized my context skill into a company-wide tubi-context skill and opened a PR to Tubi's shared Claude skills repo. Published a catalog of every Claude MCP, skill and plugin at Tubi, then a lighter one when the first proved too dense. Ran AI trainings and 1-on-1s.
The plumbing
Wired 11 tools into the setup, mostly by debugging real breakage. I built an 8-file memory system that recorded locked decisions with what would unlock them. Memory writes were proposed with a diff. Never automatic. When a scheduled job runner turned out to have failed silently for weeks, I retired it for three pull-based Claude Code slash commands.
Kept score on the AI
A running list of what the AI got wrong. Nine items, among them sub-task hierarchy it silently dropped and a Slack bridge that returned zero results. I also tested which permission primitives actually enforced, found one that didn't, and re-probed to correct my own earlier hedge.
Discovery, handed to an agent
Built team-onboarder, a read-only agent that audited how a team worked and drafted its migration config. The team-lead conversation stayed in the playbook.
03
Tooling & Platform Governance
The systems teams work inside. Jira for a whole company; the field architecture, the permission schemes, the licenses. And the debugging.
TubiSr. Product Ops Manager
Ran Jira as a platform
Owned Jira for all of Tubi, 400+ employees: the tech org on company-managed projects, the business orgs on bespoke schemes. Onboarded four orgs. Wrote the audits and build guides, designed the field, form and automation architecture behind new team spaces, wired Slack into the customer-experience issue intake, ran licenses and permissions, and was the escalation path to Fox's global admins.
The accounting side of Jira
Documented the Jira rules that make eight figures a year of engineering labor capitalizable. The hierarchy went initiative to epic to story. Work types drew the line finance needed between new development and maintenance.
Didn't pick them, made them work
Implemented and governed Statsig, PagerDuty and Resolve AI as internal tools. Configuration, rollout, enablement, documentation. On Statsig I also reviewed conflicts between teams' tests, and the experiments themselves stayed with their teams. When PagerDuty license usage ran past expectations, I ran an audit to find seats we could reclaim.
Fit the structure to how teams worked
Audited how each of the later teams actually worked before building its space. Two groups ended up as work types inside Content Ops. A third got a request form.
Sling TVEng Leader, Product Ops & SRE
PagerDuty, the first time
Sling's PagerDuty admin as well: the on-call platform, the SRE reporting, the process improvements around it. I owned the same platform again at Tubi.
Sling TVSr. Manager, Care Tools
Two picked, two inherited
Ran the care tooling at Sling. Selected and implemented Freshdesk for the agents and Sprinklr for social care. Adobe and the Pypestream chatbot came with the job, and I moved the help site onto Adobe.
Tickets you could trust
Rebuilt Sling's support ticketing so its data was accurate enough to analyze and triage against.
04
Customer Experience
Where I started, and what still grounds the rest. Contact centers, self-service, voice of customer, and for a year and a half, the whole org.
Sling TVHead of CX
Led all of CX
Cut the cost of customer care by 78%. 400+ business initiatives, 300+ app and web updates, 200+ bug fixes and a re-focus on self-help. Agent-involved contact fell from over 80% to roughly 20%, while satisfaction went up 25%. About 70% of the cut came from the chatbot, 20% from the contacts agents still handled, 10% from the help site.
Eighteen people: three managers, fifteen ICs
One of the senior leaders in a total rebuild of Sling's org structure. The org I led came out of it. Product ops, Digital Care Products, the NOC and voice of customer all sat inside it.
Sling TVSr. Manager, Care Tools
Redesigned around user feedback
Rebuilt the help site, about 250 articles. A 'did this help?' widget on every page plus a CSAT survey showed that people couldn't find what they were looking for. So we replaced the top-center search box with categories. Help-site CSAT rose 10%+. Click-through to the chatbot dropped a long way, and contact volume fell too.
Sling TVHead of CX
What customers were living through, where decisions got made
Integrated real-time customer experience data into the core business processes. Conviva gave us playback quality. Voice of customer, contact volume, NPS and CSAT came through Kibana.
Co-owned the CX budget
With business development. I also ran many of the vendor relationships, and partnered with Dish's care organization on agent staffing and outsourced-partner performance.
Helped with the planning
Leadership retreats, OKRs, talent planning, metrics and targets, in this job and the care-tools job before it. I care about the internal engagement survey more than I probably should. We consistently beat industry and company benchmarks on it.
Cart, offers and billing
Some of the 300+ app and web updates went into those flows, aimed at the contacts they generated and at conversion.
Sling TVSr. Manager, Care Tools
The money questions
Extended the self-service on both the chatbot and the help site into billing and offers.
Sling TV
One job that kept getting bigger
Build it, staff it, hand it off, take the next thing. I built UAT and rebuilt the NOC, then took product ops and the NOC, then care tools, then the whole care org. Then engineering.
Led managers for five-plus years
The NOC manager from around 2018, three managers in the care org, then the SRE lead in 2022–24. I hired dozens of people at Sling and ran hundreds of interviews, for teams I built from scratch and teams I inherited. Promotions were a mix. Planned, battlefield, and the kind a restructure forces.
Dish Network
Went to the floor
Traveled to contact centers in the U.S., Mexico and the Philippines to hear what agents and customers were running into, fix bugs, and help steer the agent tools and products.
Dish NetworkAnalyst, Customer Support
Before streaming
Started at Dish Network's headquarters, working a lot with the contact centers. Moved into technical support project management two years later.
05
Change & Migration
Moving people between systems is mostly not a technical problem. At Tubi it meant a hard vendor cutoff, teams that modeled work differently, and the accounting afterward.
TubiSr. Product Ops Manager
Mine alone
Ran the migration off Shortcut, a retiring work-tracking system, on the assumption that anything not exported before October 1, 2026 was gone for good. The tech org went first, 20 product and engineering teams in one program. Then team by team. I verified 5,700+ tickets from those later moves against the migration logs.
Adding a team meant editing a config file
Built the migration engine with Claude Code: 44 scripts and ~14,650 lines of Python. When Creative Studio's destination changed to Airtable, I wrote a second destination adapter and left extraction, normalization and provenance alone.
Sling TVManager II, Product Ops & NOC
Unwinding the framework
I helped lead the move off SAFe, to empowered teams releasing against SLOs. Quality dipped for a while, then recovered. Delivery speed finished about 18% up.
TubiSr. Product Ops Manager
An intake gone quiet
Found a failed cutover. A customer-experience intake had stalled for eight months, with 153 tickets on track to disappear at the cutoff. I raised it with a named owner.
Nobody woke up to a notification flood
Each team-by-team move went in phases: a Jira preview, a quiet first load, then a people sync. Runs were resumable, with written rollback criteria. Every full run waited on a validation gate, like a 20-ticket test batch for Content Ops or five tickets first for Creative Studio.
Reversible first
Split the Ad Ops build in two. Track A was everything reversible; Track B was 32 tenant-wide custom fields you can't take back. I held B while a tool decision was pending, and when that decision slipped, I went ahead with A.
Built to be handed off
Delivered the Content Ops Jira training in live sessions on April 27, 2026, then fixed filter sharing when people couldn't see their boards. Wrote the team migration playbook and a training deck template so a successor could run the tool.
What actually cut over
I copied four teams' data: DSA, Growth Marketing and Content Ops (1,129 issues) into Jira, Creative Studio into Airtable. Then I reported what that did and didn't mean. Only DSA fully cut over. Shortcut was never locked for writes, so data kept changing after it was copied, and my status report named that as the root cause.
Remaining scope, recounted
Changed the liveness measure from the date a ticket was last updated to the date it was created. Seven to nine teams of remaining scope became three.
Sling TVManager, UAT
Teaching the new process
When Sling adopted SAFe, I was the process lead and one of the implementation leads. I also trained teams on it.
Release cadence
Moved clients and platforms off quarterly releases. Monthly major releases replaced them, with smaller releases and fixes in between.
06
Incident Command
Seven years on call. I've run the response and the practice around it, and written the after-action analysis for leadership.
TubiSr. Product Ops Manager
Owned the incident program
Set a goal in fall 2025 to bring resolution time down and hit it — moved on-call from Opsgenie to PagerDuty and trained teams on the process. MTTR down 15%. Ran the monthly SEV reviews and tracked corrective actions in Jira, against a target of closing 90% of P0/P1 actions within 7 days.
And its leadership reporting
Wrote the weekly incident recaps and impact analyses for the executive Viewer Growth Review, sizing viewership impact with the team's accepted method.
Put a number on a gap
Found that only about 40% of SEV0-2 postmortems had a matching incident in Datadog. Also ran incident volume half-over-half, then re-ran it outlier-controlled when the raw number proved skewed.
Sling TVEng Leader, Product Ops & SRE
No runbooks. No baselines.
Rebuilt incident and critical-event management at Sling with the executive team behind it. The on-call I inherited paged too many people. Led 120+ incident responses and 30+ critical-event calls myself. MTTD fell 68%, MTTR 28%.
Then the basics
Put in on-call tooling, automated alerting and a structured RCA and retro process. Deployment frequency rose 15% while reliability improved.
Sling TVManager II, Product Ops & NOC
Where I first learned to be an incident commander
I ran the NOC. It monitored every production release and triaged emerging issues with the context engineers needed. I developed its incident management practices and hired its manager, who then ran a process update and automation upgrade.
Sling TVManager, UAT
Two teams at once
Took on and rebuilt the NOC while still managing UAT. It didn't come with a new title.
07
Live Sports & Events
Hundreds of live events at Sling. Then Tubi, where I was the live-TV veteran in the room as the company learned to stream live.
TubiSr. Product Ops Manager
World Cup matches, F1 races, football games
Ground Control Lead for Tubi's live events, tracking real-time issues during the broadcast. I built the tracker as a Claude skill, linked to Datadog dashboards, delivered by DM only and hardened against prompt injection, with columns matched to the runbook's. Created the Jira space for live-event work, too.
Seen this one before
Ahead of launch, I caught game scores showing on replay tiles, changed the view and added a QA case. I'd seen the same bug at Sling.
Sling TVManager II, Product Ops & NOC
Nights and weekends
Ran the NOC and the live-event war room, jointly with an executive, through hundreds of live events: football, basketball, PPV, and early-Sunday-morning cricket.
TubiSr. Product Ops Manager
Time-behind-live testing
It had gone 6+ months without a run. I got it restarted and running more often, since lag varies by CDN and platform.
Moved live-event support onto standard incident practice
I ran post-event retros as short decision meetings. A Claude retro bot gathered the basic insights first, and the follow-ups were ranked by impact and timing.
Who loses the stream
Pushed through a concurrency rule so that when stream limits hit, viewers already watching weren't the ones dropped.
Sling TVHead of CX
PPV boxing
Ran the customer side: staffing with Dish's care organization, care-tool messaging, help-site banners, internal and external status pages, and the on-channel slate.
Sling TVSr. Manager, Care Tools
A/B-tested the stream itself
The tests covered CDN selection, the ads experience, and a built-in playback delay that preloads to avoid buffering.
Sling TV
Playbook owner and requirements writer
Incident commander for live events. That covered CDN cutovers, failsafes, detection and monitoring, and the dashboards. The process, documentation, retros and tracking I built for them became the standard. On the product side, I wrote requirements for score hubs and real-time excitement ratings.
The Thanksgiving CDN cutover
I called it during an ad break in a live NFL game, working from the playbook.
08
Ad Tech
I know ad tech from the operating side: testing and monitoring ads, and hearing from viewers when one failed. Most of that was at Sling.
TubiSr. Product Ops Manager
Jira for Ad Ops and Ad Sales
Wrote the audits, build guides and intake designs for both. Ad Ops was the largest team left to move.
When ad serving broke
Those incidents went into the recaps and impact analyses I wrote, same as any other. Outside incidents, I helped test contextual ads.
Sling TVHead of CX
Heard it from customers
Kept ad-based feedback and ad complaints a top priority across the customer experience org I led.
Sling TVManager, UAT
Tested ads, then monitored them
UAT covered ads and live events. Lots of ad hoc testing and feedback on ads, and learning the different ad types and strategies. When I took on the NOC, I overhauled its ad monitoring.
Sling TV
Sling had no dedicated ads PM
I was a lead voice on ads and their effect on customers, and joined leadership discussions on ad volume, repeated ads and frequency.
Sling's first ad alerting, in Conviva
I built it to catch ad spikes and drops, ad lengths and hang-ups. Freshdesk and Sprinklr got ad alerting too. I also worked with FreeWheel and rebuilt Sling's ad reporting entirely.
Changed the incident process
Ad incidents stepped up a priority level, so an ad P3 was handled as a P2.
The workaround
An SSAI issue once grew into a contractual problem, and Sling ran pre-rolls until the underlying fix was in.
The conventional version
Same career, ordered the usual way: years, titles and teams, with the top bullets for each role. Everything above is the long version.
No direct reports — a deliberate move to a top-tier tech company where everyone up to SVP builds and does IC work, and you manage outcomes, not people. What I held: Jira company-wide, the incident program, the migration off Shortcut, the creator program in lieu of a dedicated PM, and live-event support.
- Acted as PM for the creator program in lieu of a dedicated PM, and flagged the missing role in writing as a staffing gap that tooling couldn't close, so it wouldn't be inherited silently. I carried the half-year plan overview and the weekly product and cross-functional syncs.
- Owned Jira for all of Tubi, 400+ employees: the tech org on company-managed projects, the business orgs on bespoke schemes. Onboarded four orgs. Wrote the audits and build guides, designed the field, form and automation architecture behind new team spaces, wired Slack into the customer-experience issue intake, ran licenses and permissions, and was the escalation path to Fox's global admins.
- Ran the migration off Shortcut, a retiring work-tracking system, on the assumption that anything not exported before October 1, 2026 was gone for good. The tech org went first, 20 product and engineering teams in one program. Then team by team. I verified 5,700+ tickets from those later moves against the migration logs.
- Set a goal in fall 2025 to bring resolution time down and hit it — moved on-call from Opsgenie to PagerDuty and trained teams on the process. MTTR down 15%. Ran the monthly SEV reviews and tracked corrective actions in Jira, against a target of closing 90% of P0/P1 actions within 7 days.
Ten people: a seven-person 24/7 SRE team in Utah, a two-person QA team in India, and a principal engineer on Critical Problem Resolution — a two-person function, that engineer and me.
- Built the product-ops practice for engineering at Sling. CX and quality-of-experience data surfaced the likely bugs; we confirmed them, sized scope and risk, and got them prioritized against the live sprint schedule. 270+ production fixes shipped that way.
- Rebuilt incident and critical-event management at Sling with the executive team behind it. The on-call I inherited paged too many people. Led 120+ incident responses and 30+ critical-event calls myself. MTTD fell 68%, MTTR 28%.
The whole care org, three managers and fifteen ICs: the VoC program manager, the tools PMs, UAT, plus product ops and the NOC I already had.
- Cut the cost of customer care by 78%. 400+ business initiatives, 300+ app and web updates, 200+ bug fixes and a re-focus on self-help. Agent-involved contact fell from over 80% to roughly 20%, while satisfaction went up 25%. About 70% of the cut came from the chatbot, 20% from the contacts agents still handled, 10% from the help site.
The care-suite PMs (chatbot, help site, agent tools for chat and voice), product ops on the care side, and the NOC.
- Managed the team of product managers over the care suite: the chatbot, the help site, and agent tools for chat and voice. I owned the A/B testing program too. Help-bot flows, website designs, call-routing patterns.
- Inherited Pypestream, Sling's customer-facing chatbot. My team rebuilt it from the ground up while the vendor was still iterating on the product underneath. We worked the most common customer workflows hard, and containment passed 90%.
- Rebuilt the help site, about 250 articles. A 'did this help?' widget on every page plus a CSAT survey showed that people couldn't find what they were looking for. So we replaced the top-center search box with categories. Help-site CSAT rose 10%+. Click-through to the chatbot dropped a long way, and contact volume fell too.
Product ops (Sling's service stability team) and the NOC, whose manager I hired.
- Established the Product Operations team at Sling — the people who sat with product managers to set KPIs and SLOs, then tracked them.
- Ran the NOC and the live-event war room, jointly with an executive, through hundreds of live events: football, basketball, PPV, and early-Sunday-morning cricket.
- I helped lead the move off SAFe, to empowered teams releasing against SLOs. Quality dipped for a while, then recovered. Delivery speed finished about 18% up.
Built from scratch: four people, then two more absorbed from another team in a restructure.
- Created Sling's first user-based testing team. The scripts I inherited tested what the PRD said a feature did. Not what a customer would actually do with it. A new suite written around user behavior caught thousands of issues before release. The team tested cart, offer and billing releases too.
- Took on and rebuilt the NOC while still managing UAT. It didn't come with a new title.
Education. University of Minnesota, Carlson School of Management — BS Business, Marketing; minor in Management Information Systems (2013)
How I work
Pull, not push. The people on my team should be pulling work off my plate, not pushing work onto it — which means they make their own calls and I'm the backboard, there for guidance when they want it. I learned to manage people, and to work well with them, by building trust.
On the side
Games and apps built on my own time.
Built by directing Claude, for my six-team fantasy NBA league, where every pick has to be rated 85 or under in NBA 2K. Live since October 14, 2026, on Cloudflare Pages.
Flutter and Dart, built with Claude Code and a seven-agent setup.
Details when it ships.
Get in touch
- linkedin.com/in/mike-danley
- Resume
- On request — just askmichael.r.danley@gmail.com · subject: Resume request