● OTT/CTV Portfolio · Broadcasting
YASH SHAH
Senior PM · OTT / CTV · AI Enthusiast
Home Work Skills Experience Blog Contact
AI Enthusiast · Streaming PM · OTT/CTV · Ad Tech

Hi, I'm Yash.

LIVE · OPEN TO ROLES

I am a product manager who likes turning messy ideas into products people can actually use.

My work has mostly been in streaming, ad tech, live experiences, AI-assisted workflows, and platform products, but I care more about the problem than the industry label.
I like owning the unclear parts, working closely with engineering, and making complicated products feel simple for the people using them.

▶ View My Work Read My Blog
AI Enthusiast Streaming AI CTV Discovery OTT Automation Ad Tech
Let's Connect LinkedIn
10+ yrs
Streaming
12+
Platforms
45%
Rev Growth
120%
Engagement
40+
Team Led
Yash Shah
Current Focus

Product experiences I am building toward.

AI Discovery
AI : Short-Form Discovery
AI discovery that turns clips into viewing.
Trailers, teasers, sports highlights, movie moments, series hooks, and live-channel clips can help viewers decide faster. The goal is more interaction, stronger engagement uplift, and longer viewing time.
AI discovery ranks sports, movies, series, and live moments by intent Clip to full movie, episode, live event, or channel Measure engagement uplift, interactions, starts, and viewing time
AI Sports Highlight
Match point, ranked for intent.
+42% interaction signal
Watch live
Movie Moment
The scene that makes someone press play.
Predicted completion: high
Start movie
Series Teaser
Hook the episode without spoiling it.
Next-best title match
Open episode
AI Sports Highlight
Match point, ranked for intent.
+42% interaction signal
Watch live
Movie Moment
The scene that makes someone press play.
Predicted completion: high
Start movie
Series Teaser
Hook the episode without spoiling it.
Next-best title match
Open episode
←
Added to watchlist
Opened from clip
Movie detail
The feed jumps into the title page, adds context, then lets the viewer play or save it.
PlayWatchlist
AI DiscoverRank clips by sport, genre, mood, completion, and watch history.
EngageCapture likes, skips, rewatches, shares, and click-through intent.
WatchMove from short clip to full viewing with one obvious action.
+120%Interaction
+38%Starts
+24%View Time
AI Workflow
AI Content Onboarding
From raw package to live publishing.
The operator enters a movie, series, or sports title and AI builds the onboarding package: metadata, poster assets, trailers, teasers, captions, ad breakpoints, and review-ready publishing data.
For movies and series, AI fetches metadata, posters, trailer assets, teaser assets, captions, and artwork For sports, AI creates highlights, clips, moments, metadata, thumbnails, and publishable short-form assets AI recommends ad breakpoints and operators approve exceptions before publishing live
Enter title or sports event
Live Sports Event
1
IdentifyMatch movie, series, episode, live game, or sports event.
2
FetchPull metadata, artwork, poster, trailer, teaser, and caption assets.
3
CreateBuild highlights, clips, thumbnails, AI ad breakpoints, and promos.
4
PublishOperator reviews exceptions, schedules, and pushes live.
MetadataTitle, teams, cast, genre, synopsis, EPG, tags
PostersHero art, key art, thumbnails, vertical covers
VideoTrailer, teaser, highlights, clips, moments
AdsAI ad breakpoints, SCTE-35 markers, pod cues
Poster assets
Trailer + teaser
Sports clips
Ad breakpoints
Ingest
AI Enrich
Operator QA
Schedule
Publish Live
Sports Multiview
CTV Live Sports
Sports multiview plus live polling.
For live sports, I like experiences that make the TV feel alive without becoming complicated. Four video views can support basketball, baseball, football, soccer, highlights, stats, and alternate feeds while polling gives fans a fast way to react.
Multiple live video tiles for game, highlight, stats, or alternate feed Remote-friendly live poll overlay for fast participation Engagement loops that still respect the lean-back CTV experience
SOCCER · LIVE
BASEBALL · LIVE
FOOTBALL · RED ZONE
BASKETBALL · STATS
Live Poll
Who wins the next run?
AI in Practice

Not just ChatGPT prompts.
Actual AI flows in production.

Every capability below is something I have shipped, defined, or built a product around. Real pipelines. Real models. Real outcomes.

🤖
Content Intelligence
AI Metadata Pipeline
Multimodal model reads video frames, audio, and transcript to auto-generate genre tags, mood tags, content warnings, and searchable keywords. Replaces a 4-hour manual workflow with a 90-minute automated one.
Multimodal LLMVideo AnalysisNLPAuto-tagging
↑ 60% faster content ops
🎭
Discovery Ranking
Intent-Based Clip Ranking
Ranking model scores short-form clips by predicted watch-through rate, genre match, viewer history, time of day, and recency. Rankings update in real time as new engagement signals arrive from the feed.
Ranking ModelReal-Time SignalsEngagement ScoringFeature Store
↑ 42% interaction signal on ranked clips
🎵
Captioning & Localisation
AI Captions in 70+ Languages
ASR model generates word-level timed captions. Neural machine translation covers 70+ languages with speaker diarisation and style-aware formatting. Output in SRT, VTT, and TTML for direct SSAI and EPG integration.
Whisper / ASRNMT TranslationSpeaker DiarisationSRT / VTT / TTML
70+ languages · Word-level timestamps
🎬
Sports AI
Automated Highlight Generation
Computer vision detects match events — goals, touchdowns, key plays — from live or VOD video. Each moment becomes a timestamped clip with generated title, thumbnail, and engagement prediction, published to the discovery feed within minutes.
Computer VisionEvent DetectionAuto-ClippingThumbnail AI
↑ 38% more content starts from highlight feed
📈
Ad Intelligence
AI Ad Breakpoint Detection
Model analyses audio and visual cues to identify natural ad break positions. Output is a confidence-scored list of breakpoints in SCTE-35 format. High confidence breaks go live automatically. Low confidence ones go to human review.
Audio AnalysisScene DetectionSCTE-35 OutputHuman-in-the-loop
↓ 50% manual marking effort
🎯
Personalisation
Next-Best-Title Recommendation
Collaborative filtering model trained on watch history, completion rates, search behaviour, and genre affinity. Serves personalised content rails across home screen, discovery feed, and post-watch end card. Model retrained daily. A/B tested via SDUI.
Collaborative FilteringEmbedding ModelA/B ExperimentationSDUI Integration
Personalised for every viewer · Daily retraining
Selected Work

Products I have taken
from problem to launch.

A few examples of how I think, what I owned, and where the work moved the business.

01/03
Short-Form Video · Discovery · Monetization
Short-Form OTT Discovery for CTV and Mobile
Scroll-to-watch journeys · In-feed ads · 60% faster content ops
The Problem
Discovery was doing too much heavy lifting with posters. Operators had great trailers, clips, and live moments, but the app experience still pushed viewers through static rails.
What I Did
I shaped a short-form discovery surface for OTT. Viewers could move through trailers, teasers, clips, and live-channel moments, then jump into the full movie, series, episode, channel, or clip. I also worked through in-feed ad insertion so discovery could create revenue without getting in the way.
The Outcome
Browsing became a product surface, not just a menu. It gave operators a modern way to promote content, create intent, and open a new ad inventory model.
60%
Ops Cut
New
In-Feed Ad Inventory
CTV
+ Mobile Discovery
Tap
To Full Watch
📈 Also shipped: Ads Between EPG Slots
In the Live EPG view, I inserted ad inventory between programme slots — the moment a viewer is deciding what to watch next. SSAI-served 15s ads run between grid rows during natural pause moments, opening a net-new ad position that does not interrupt playback.
SSAIEPG Grid AdsTransitional InventoryNon-Interruptive
02/03
Live Engagement · Sports · News
Live Polling and WatchParty for CTV
↑ 120% in-app interactions · New live engagement model
The Problem
Live viewers were engaged, but the product was quiet. Sports and news created strong opinions in the room, then pushed people to other apps to react.
What I Did
I designed Live Polls and WatchParty for the TV remote. The hard part was keeping interaction lightweight on a D-pad. I also helped define the producer CMS, real-time poll flow, and A/B tests across sports and news.
The Outcome
In-app interactions grew 120%. The model helped live sessions feel more participatory and became a strong demo story at NAB and IBC.
120%
Interactions
↑
Session Length
↓
Live Churn
2+
Verticals
03/03
SDUI · Platform · Monetization
One Product Across Twelve Screens
↑ 45% revenue growth · 12+ platforms unified · 70% CDN gain
The Problem
Every platform moved at a different speed. A simple UI change could turn into separate releases for Roku, Fire TV, Apple TV, Android TV, Samsung, LG, mobile, and web. The ad stack also needed cleaner beacon tracking and stronger integrations.
What I Did
I helped make SDUI the product standard. I owned the shared design system and worked through SCTE-35 pod logic, beacon tracking, Google DAI, Yospace, and Freewheel across AVOD, SVOD, TVOD, and PPV use cases.
The Outcome
Revenue grew 45% over 18 months. Experiments moved faster, CDN efficiency improved 70%, and the platform became easier to sell and operate.
45%
Revenue Growth
70%
CDN Efficiency
12+
Platforms
Days
Experiment Cycle
How I Think

The product instincts
I keep coming back to.

01 / 04
Start with the viewer.
Catalogs, rights, and monetization matter, but the viewer only cares whether the experience is worth their time. That is where I start.
02 / 04
Use data, then ask better questions.
A dashboard can tell me where something changed. It cannot tell me the whole story. I like pairing analytics with real conversations and product judgment.
03 / 04
Make the complex parts feel normal.
Most viewers will never think about SSAI, SDUI, CDN logic, or device fragmentation. That is fine. The product is working when they do not have to.
04 / 04
Ship with engineering, not around it.
The best product decisions usually come from tight conversations with engineering, design, QA, ad ops, and customers. I try to keep those loops close.
Impact

The numbers I can
stand behind.

01/08
0%
Revenue Growth
Monetization redesign across ad pod logic, integrations, and delivery performance.
02/08
0%
Engagement Lift
Live Polls and WatchParty for sports and news on CTV.
03/08
0%
Faster Ops
AI-assisted metadata, captions, breakpoints, clips, trailers, and discovery workflows.
04/08
0%
CDN Efficiency
VOD delivery redesign with cleaner caching and customer-visible stability.
05/08
0%
Less Scheduling Work
Browser-based EPG, drag-and-drop planning, analytics, and multi-language support.
06/08
0+
Team Led
Product, engineering, design, QA, DevOps, sales, and ad ops across regions.
07/08
0+
Platforms Shipped
Roku, Fire TV, Apple TV, Android TV, Samsung, webOS, Vidaa, Vega, Whale, iOS, Android, and web.
08/08
0+
Years in Streaming
From Business Analyst to Senior Product Manager in OTT and CTV.
Process & Craft

How I move from
problem to ship.

01
Find the real user pain.
I use data to spot the signal, then talk to customers, operators, and internal teams to understand what is actually happening.
02
Write the problem clearly.
Before screens or tickets, I want everyone aligned on the user, the business goal, the constraints, and what success will look like.
03
Ship in useful pieces.
I like small launches with clean measurement. Poll placement, ad pod rules, EPG flows, and discovery changes all get better when the team learns quickly.
04
Stay close to engineering.
I read architecture docs, write SQL, review API tradeoffs, and keep product decisions grounded in how the system really works.
05
Use AI where it earns trust.
AI is great for compression, tagging, summarizing, and pattern work. I still keep judgment, prioritization, and customer empathy human.
Discovery & Research
User InterviewsAnalyticsCompetitive TeardownsCustomer Calls
Planning
PRDs / MRDsOKRsStory MappingJIRAConfluence
Data & Experimentation
SQL / BigQueryLookerTableauA/B Testing
Design
FigmaMiroWireframesSDUI Mockups
AI
Claude AIChatGPTNotion AI
Education
M.S. Computer Science
Cal State San Marcos · GPA 3.7
This technical foundation helps me contribute in system design conversations, not just listen from the side.
Platform Coverage

Shipped on
every screen.

Roku
Roku
Fire TV
Apple TV
Apple TV
Android TV
Samsung
Samsung Tizen
LG webOS
LG webOS
Vidaa OS
Vega OS
Whale OS
iOS
iOS
Android
Android
Web
Experience

A decade in
OTT and CTV.

OCT 2022 — PRESENT · WISEDV INC., SAN DIEGO, CA
Senior Product Manager
OTT/CTV technology platform serving tier-1 streaming operators globally
  • Own OTT product strategy and execution across 12+ platforms including Roku, Fire TV, Apple TV, Android TV, Samsung, webOS, Vidaa, Vega, Whale, iOS, Android, and web.
  • Built Sports WatchParty and Live Polls, driving a 120% lift in engagement and extending the model across more content verticals.
  • Led SSAI, CSAI, ad stack, and CDN improvements that contributed to 45% revenue and viewership growth over 18 months.
  • Shaped short-form OTT discovery with trailers, teasers, clips, live-channel moments, scroll-based ads, and click-through paths into full viewing.
  • Used AI-assisted workflows for tagging, captions, breakpoint detection, and metadata enrichment, cutting operations turnaround by 60%.
  • Helped standardize SDUI and a shared design system across 12+ platforms while staying close to front-end engineering decisions.
  • Led a 40+ person global cross-functional team and represented WiseDV at NAB, IBC, CES, and Streaming Media Show.
FEB 2021 — OCT 2022 · WISEDV INC., SAN DIEGO, CA
Product Owner
  • Launched cloud-based FAST channel playout with 4K, SSAI, SRT, HLS, RTMP, and 24×7 broadcast-grade operations.
  • Built a browser-based drag-and-drop EPG that reduced manual scheduling effort by 50%.
MAY 2020 — AUG 2020 · DVEO, SAN DIEGO, CA
Integration Specialist (Contract)
  • Re-architected VOD delivery, improved CDN efficiency by 70%, and kept the migration invisible to customers.
JUL 2014 — JUN 2020 · WISEDV ELECTRONICS INDIA PVT. LTD., VADODARA
Business Analyst
  • Owned requirements for FAST playout, SSAI, multi-platform delivery, CMS workflows, BRDs, FRDs, API docs, and NAB/IBC demos.
Skills

The areas I work
in every week.

Streaming & Video
Live StreamingVODFAST4K/HDRABRHLSDASHSRTDVRMulti-DRMCDN
Ad Tech & Monetization
SSAI/CSAIAVODSVODTVODGoogle DAIYospaceFreewheelSCTE-35Roku RAF
Discovery & Engagement
ML PersonalizationA/B TestingLive PollingWatchPartyShort-Form VideoScroll Feed UXIn-Feed AdsAI Ranking
Architecture
Server-Driven UIDesign SystemsCTV UXEPGCross-PlatformFE Engineering
Product Craft
PRDs/MRDsOKRsKPI FrameworksAgile/ScrumGTMStakeholder Mgmt
AI for Streaming
AI DiscoveryAI RankingMetadata AutomationCaption WorkflowsBreakpoint DetectionClaude AIChatGPTSQL / BigQuery
Let's Talk

Looking for a PM
who understands
streaming systems?

I am open to senior PM roles in streaming, OTT/CTV, media tech, and platform products. I am based in San Diego and can work remote, hybrid, or in person.

Available Now · Remote · Hybrid · In-Person
yash_121@yahoo.com (858) 371-8919 LinkedIn Profile
San Diego, CA
✓
Message Received

Thanks! I'll get back to you within 24 hours.