Vasista Sandeep, AVP Program Management (Tech & Product)

Program & Product Leadership · OTT · FinTech · Retail

Building platforms that scale to 30M+ users.

I turn ambitious roadmaps into resilient, high-concurrency products, aligning engineering, product, and delivery so the platform stays fast when the audience shows up all at once.

Years Technology Leadership
0+Years Technology Leadership
Peak Concurrent Users
0M+Peak Concurrent Users
Peak Requests / Sec (live events)
0KPeak Requests / Sec (live events)
System Availability
0.00%System Availability
Publishing Latency Reduction
0%Publishing Latency Reduction

  1. Sony Pictures Networks India

    OTT / Streaming

    AVP, Program Management (Tech & Product)

    Mar 2026 – Present

    Governing strategic delivery and the technical roadmap for the global SonyLIV OTT platform, serving 30M+ concurrent users.

    • Organizational Leadership & Scale: Direct the executive engineering portfolio and a cross-functional 'Team of Teams' to sustain 99.99% system availability during extreme live-traffic events (250K RPS).
    • Personalization Strategy: Spearhead the technical strategy and scaling of personalization engines, driving complex cross-team integrations to deliver highly tailored end-user experiences.
    • Cross-Functional Governance: Act as the central bridge between Product and Engineering leadership, ensuring technical delivery aligns with overarching business priorities.

    Lead Program Manager (Tech & Product)

    May 2021 – Mar 2026
    • Strategic OPEX & Build vs. Buy: Managed the multi-million-dollar budget and delivery for 'Bolt' (proprietary CMS), replacing legacy vendors to achieve a 40% reduction in publishing latency and major operational savings.
    • Data Governance & Scale: Led cross-functional squads (Legal, InfoSec, Engineering) to define technical roadmaps for US/Europe expansion, implementing robust GDPR/PII compliance frameworks.
    • Innovation Strategy: Steered the 'Mobile Web Revamp' and 'Shorts' video initiatives, driving decisions that optimized rendering performance and increased ad-inventory monetization.
    • Architectural Modernization: Drove the modularization of a legacy monolithic backend, decoupling and scaling dedicated microservices for Playback, Listing, and User Subscription Management.

  2. Walmart Global Tech

    FinTech / Retail

    Technical Program Manager

    Dec 2019 – May 2021
    • FinTech Architecture & Operational Excellence: Led the engineering roadmap for the 'Post Payment Audit' system, partnering with Staff Engineers to redesign data pipelines for automated anomaly detection while championing DevOps/CI-CD practices — recovering multi-millions annually.
    • Retail Commerce & Platform Optimization: Governed cross-functional delivery of in-store financial service platforms, minimizing transaction latency, accelerating release cycles, and elevating the Point-of-Sale (POS) experience.

  3. Conduent Inc.

    AI / Analytics

    Business Process Manager

    Mar 2019 – Dec 2019
    • AI/ML Strategy: Spearheaded development of NLP-based machine-learning models for customer sentiment analysis, enabling data-driven strategies to reduce churn.
    • Data Visualization: Delivered executive dashboards (Power BI) visualizing revenue trends and operational efficiency, directly informing quarterly business reviews (QBRs).

  4. Intel Corporation

    Data / Platforms

    Senior Business Analyst

    Jul 2016 – Mar 2019
    • Platform Integration: Managed vendor relationships and delivered an integration platform handling 10,000+ daily transactions, generating $10M in business value.
    • Big Data Analytics: Led data-analysis initiatives using Impala and Hive, applying predictive modeling to optimize supply-chain decisions.

  5. Early Career

    Business Analysis

    Business Analyst — Odessa Technologies

    Nov 2015 – Jun 2016
    • Business analysis for enterprise software delivery.

    Business Analyst — Accenture

    Jul 2012 – Nov 2015
    • Business analysis and delivery across enterprise programs.

  • Cricket

    • India Bilateral Tours vs England, Australia, and South Africa
    • ACC Asia Cup
    • The Hundred

    0 to 250K RPS spikes in under 60 seconds

  • Football

    • UEFA Champions League Final
    • UEFA Europa League Final
    • UEFA Euro
    • FA Cup

    zero-buffer 1080p and 4K adaptive bitrate

  • Multi-Sport & Tennis Slams

    • Tokyo Olympics
    • Asian Games
    • Australian Open
    • Roland Garros
    • US Open

    15+ simultaneous live feeds with multi-language metadata

  • Entertainment & Combat Sports

    • WWE WrestleMania
    • WWE SummerSlam
    • UFC Fight Nights
    • KBC Play Along real-time interactive low-latency sync

    real-time interactive low-latency sync

  1. Layer 1: Client & Edge

    • Player SDKs
    • Multi-CDN Steering
    • Token-Bucket Rate Limiting

  2. Layer 2: API Gateway & SSAI Proxy

    • Manifest manipulation
    • SCTE-35 marker detection
    • Auth

  3. Layer 3: Decoupled Domain Microservices

    • Playback
    • Listing
    • Subscription
    • XDR Continue Watching

  4. Layer 4: Persistence & Caching

    • Multi-cluster Redis pre-warming
    • Database-per-service
    • Kafka event bus
  • Circuit Breakers
  • Token-Bucket Rate Limiting
  • Database-per-service isolation
  • Multi-CDN Origin Shielding
  • Redis pre-warming

A worked example of the cost lever behind full-stack observability. Drag the slider to set peak viewer concurrency, switch the sampling strategy, and watch the estimated monthly telemetry bill respond in real time. Every figure below is computed from the assumptions noted under the calculator—nothing is hard-coded.

30,000,000
5,000,00030,000,000

Sampling Strategy

Telemetry Ingested

150,000,000spans / sec

Volume of trace data your APM vendor bills for.

Estimated Monthly Telemetry Cost

$14,040,000

Baseline at 100% ingestion: $14,040,000 / month

Standard 100% ingestion captures everything but is the most expensive option. Switch to tail-based sampling to see the savings while keeping 99.9% MTTR fidelity.

How this estimate works

  • Each concurrent viewer emits ~5 telemetry spans/sec across instrumented hops.
  • Costs are normalized over ~40 peak live-event hours per month (not 24x7).
  • Blended APM ingest price of ~$0.65 per million spans.
  • Tail-based sampling keeps 100% of error traces + 1% of healthy traces.

Figures are illustrative order-of-magnitude estimates to demonstrate the observability cost/latency trade-off — not a vendor quote.

Toggle each card between the Before (the problem we inherited) and the After (what we shipped).

  • Monolith to Cloud-Native Microservices

    Before

    A single monolithic application where playback, listing, subscription, and cross-device resume shared one deployment, coupling release cycles and blast radius.

  • Proprietary Bolt CMS

    Before

    A third-party vendor CMS imposing licensing costs, vendor lock-in, and slow, inflexible publishing workflows.

  • GDFO Graceful Degradation and Fallback Operations

    Before

    During traffic surges, a 3-minute autoscaling provisioning lag left the platform exposed to cascading failures and full outages.

  • Multi-Region Global Footprint

    Before

    A single-region deployment that could not satisfy data residency requirements or serve global audiences with low latency.

  • SSAI/DAI Ad Monetization

    Before

    Client-side ad insertion vulnerable to ad blockers, stitching failures, and buffering that eroded ad revenue and viewer experience.

  • Unified Observability

    Before

    Fragmented, siloed logging and metrics with no distributed tracing, leaving incidents hard to diagnose and slow to resolve.

  • LivPulse

    OTT Observability

    A real-time OTT platform health and analytics dashboard surfacing streaming KPIs and system pulse in one view.

    • OTT
    • Observability
    • TypeScript
    • Dashboards
    CodeLive

  • OTT KPI Benchmarking Engine

    OTT Analytics

    A benchmarking engine that compares OTT product KPIs against reference bands to spot outliers and opportunities.

    • OTT
    • KPIs
    • Benchmarking
    • TypeScript
    CodeLive

  • OpenLoanIQ

    FinTech

    A lending-intelligence prototype that turns raw loan data into clear, decision-ready signals.

    • FinTech
    • Lending
    • Analytics
    • TypeScript
    CodeLive

  • micro-ott

    Platform Architecture

    A microservices reference for an OTT backend, modeling decoupled playback, listing, and subscription domains.

    • Microservices
    • OTT
    • Architecture
    • TypeScript
    CodeLive

  • OTTObserve Academy

    Observability / Learning

    An interactive primer that teaches OTT observability concepts (MELT, SLOs, tracing) through hands-on lessons.

    • Observability
    • MELT
    • Education
    • TypeScript
    CodeLive

  • StoryScope

    Product / Agile

    A story-scoping helper that guides teams to a faster, more consistent estimate of user-story size.

    • Product
    • Agile
    • Estimation
    • Python
    CodeLive

  • Agile Companion

    Program / Agile

    A lightweight companion for agile ceremonies and program cadences, keeping teams aligned with less overhead.

    • Agile
    • Program Management
    • Tooling
    • TypeScript
    CodeLive

  • OTT Automation Suite

    Quality / Automation

    A comprehensive OTT test-automation suite (75+ tests) with a web dashboard and cross-browser coverage.

    • Test Automation
    • OTT
    • Quality
    • TypeScript
    Code

  • Telemetry Precedes Features

    No service ships without OpenTelemetry instrumentation, SLIs, and SLO alert routing. Observability is a launch requirement, not a follow-up ticket.

  • Graceful Degradation Over Total Outage

    Survivability is engineered at the edge. Non-critical dependencies fail safely so the core experience stays available under surge and partial failure.

  • Error Budgets Dictate Velocity

    The balance of tech debt versus feature delivery is governed by SLO thresholds. Error budgets, not opinions, decide when we ship and when we harden.

  • Multi-Region by Design

    Data residency for GDPR/PII and localization are baseline requirements. Regions are a first-class design input, never a late retrofit.

  • MBA in E-Business

    2022

    Annamalai University

  • Post-Graduation in Big Data Analytics

    2018

    Illinois Institute of Technology

  • B.E. in Electronics & Instrumentation

    2012

    Visvesvaraya Technological University

  • PMP Certified

    PMI

    ID: PMP#-3195092

  • Lean Six Sigma Black Belt

    KPMG

    ID: DL122024138

  • Lean Six Sigma Green Belt

    KPMG

    ID: BG122014272

  • SAFe 5 DevOps Practitioner

    ID: 86530983-9427

  • Certified ScrumMaster (CSM)

    ID: 000594368

  • Certified Scrum Product Owner (CSPO)

    ID: 000594368

  • ITIL Certified & Practitioner

    ID: GR750213195VS

Reflections gathered from the engineering, product, and delivery leaders I've partnered with, shared by role to respect their privacy.

Head of Technology

“He is the person you want owning the hardest programs. He turned our tournament readiness into a repeatable discipline, kept engineering and SRE aligned under pressure, and when 250K RPS hit, the platform simply held.”

A leader I've partnered with

Chief Product Officer

“He speaks product and platform in the same breath. He drove our personalization roadmap - Spotlight, the trays, tray ranking - and translated deep systems work into watch-time outcomes the whole leadership team could rally behind.”

A leader I've partnered with

Chief Technology Officer

“He took on the organizational heavy lifting most people avoid - building the Team-of-Teams structure, the governance, the error-budget culture - and delivery across the org became measurably more predictable because of it.”

A leader I've partnered with

A few interactive mini-games from the world of platform leadership. No sign-up, no score-keeping server — just a quick way to see how product prioritization, incident triage, and the vocabulary of observability actually feel in practice.

Question 1 / 9🔥 Streak 00 pts
Observability

In observability, what does the acronym MELT stand for?

Say hello

Let's start a conversation

Find me here

Whether it's scaling platforms, building teams, or trading war stories from a live event, I'm always up for a good conversation.

Connect on LinkedInvasista.sandeep@gmail.com

Bengaluru, India — always happy to talk platforms

Send a Message