Available for contract engagements

Shishir Sandeep

Agentic AI Engineer · Vancouver, Canada

|
Shishir Sandeep

Engineer. Builder. Forever Red.

I'm an Agentic AI Engineer based in Vancouver, Canada, originally from Bangalore, India. I studied Electronics & Communication Engineering at Birla Institute of Technology, Mesra (2010-2014) and have spent 12+ years building production systems.

I shipped a generative AI voice agent (Sadie) into production across 15,000 restaurants handling 2M+ calls. My journey spans fintech at Oracle, BI tools at Eka, hosting infrastructure at Newfold Digital, co-founding Qwackly, supply chain orchestration at Udaan, and AI voice-agent systems at MonkeyMedia/ezCater.

Now I build agentic systems independently and take contracts. I'm running Shorts4U — a multi-agent AI video studio, and ManU4U — a sports intelligence platform for Manchester United fans. When I'm not coding, you'll find me watching the Red Devils, playing tennis, or listening to the Lex Fridman podcast.

12+
Years Exp
2M+
AI Calls Shipped
2
Startups Founded
∞
MUFC Fan Since

Experience Timeline

Click on any role to see details and tech stack.

Founder & Founding AI EngineerFounder Jan 2026 - Present
Shorts 4 U
Vancouver, Canada
  • Building and operating two agentic AI platforms end to end, across two deliberately different stacks to prove the orchestration patterns transfer
  • shorts4u.ai — Multi-agent video generation pipeline: separate agents for script, hook, storyboard, image generation, and video generation, coordinated with state carried between stages
  • Built validation checkpoints between stages so a weak hook or off-brief storyboard fails early, before it burns on image and video generation downstream
  • ManU4U — Multi-agent sports intelligence system doing tool-based reasoning over live fixtures, events, and statistics APIs with RAG pipeline over an ingested historical corpus
PythonFastAPILangchain StackLangSmithRAGVector DBAgentic AI
Sr. Software Engineer Jan 2023 - Oct 2025
MonkeyMedia Software (formerly ezCater)
Vancouver, Canada
  • Led the design and delivery of Sadie, MonkeyMedia's first production generative AI voice agent — went live across 15,000 restaurants with 2M+ calls completed
  • Built RAG-style context enrichment pipelines that pre-fetched order history, menu, and store metadata, grounding responses and satisfying 90% of user queries without escalation
  • Developed the evaluation and guardrail framework behind the agent — scoring response quality, tool selection accuracy, and hallucination rate
  • Re-engineered Authentication Server v2 on Spring Boot 3, OAuth2, OpenID Connect, and Azure Key Vault, lowering auth failure rates
  • Improved core platform reliability with OpenTelemetry instrumentation, Azure dashboards, RBAC and Managed Identity, and Terraform for infrastructure as code
Java 17Spring Boot 3Spring AIOAuth2AzureLLMRAGTerraformOpenTelemetry
Lead Product Engineer Sep 2021 - Oct 2022
Udaan.com
Bangalore, India
  • Architected backend services for capacity modelling and SLA promise accuracy across multiple fulfillment categories for 3M+ users
  • Designed breach detection and automated communication systems, reducing customer support call rates from 25% to 10%
  • Built operational dashboards with Kotlin, Redis, and React, reducing operational costs by 30%
JavaKotlinReactAzurePostgreSQLRedisCosmosDBKubernetes
Co-FounderFounder Jan 2020 - Aug 2021
Qwackly Private Limited
Bangalore, India
  • Built a celebrity-fan engagement platform from scratch — 50+ celebrities onboarded, 10,000+ user signups, 2,000 peak DAU
  • Full-stack development: React JS front-end, Spring Boot back-end, payment gateway integration, cloud hosting
  • Wore all the hats: market research, customer interviews, wireframing, investor pitch, social media marketing, Google Ads
ReactSpring BootAWSPaymentsProductMarketing
Senior Software Engineer Sep 2017 - Aug 2021
Newfold Digital (Endurance Group)
Bangalore, India
  • Developed backend services powering Linux and Windows virtualization platforms using Openstack Compute for APAC data centers
  • Owned provisioning layer revamp with containerization, caching, async queues, and centralized logging for zero-downtime deployments
  • Multiple Employee of the Month awards; Runner-Up at Global Hackathon 2019
JavaSpring BootOpenstackKafkaDockerK8sOpenshift
Software Engineer Jan 2016 - Aug 2017
Eka Software Solutions
Bangalore, India
  • Built a Power BI-like Business Intelligence tool from scratch for Commodity Analytics Cloud
  • Developed PnL web app providing real-time insight into client's realized and unrealized PnL
JavaMongoDBMySQLExtJSHighcharts
Technical Analyst Aug 2014 - Dec 2015
Oracle Financial Services Software
Bangalore, India
  • Technical support for Core Banking Platform (FCUBS) and Investment Banking/Mutual Fund Platform (FCIS)
JavaSQLPL/SQLFintech

What I'm Building

Production systems and ventures I've built.

ManU4U

ManU4U

Multi-agent sports intelligence system for Manchester United. Tool-based reasoning over live fixtures, events, and statistics APIs. RAG pipeline over an ingested historical corpus with vector embeddings and guardrails for factual grounding. Also available as an MCP for universal agent integration. manu4u.com

Agentic AIRAGVector DBPythonMCP
☕

Spring AI Blog Series

Deep-dive articles on building AI-powered Java applications using Spring AI. First article live: Navigating AI for Developers with a Java Background.

Spring AIJavaLLMsMediumTechnical Writing
🌟

Qwackly

Co-founded during the pandemic. A celebrity-fan engagement platform enabling personalized video messages. Onboarded 50+ celebrities, grew to 10K+ users with 2K peak DAU. Follow on Instagram.

ReactSpring BootAWSStartupProduct

Blog & Articles

Thoughts on AI engineering, Java, and building in public.

Agentic AI

Building ManU4U: A Multi-Agent Sports Intelligence System

How I designed tool-based reasoning and RAG pipelines to deliver real-time Manchester United insights with factual grounding.

Coming soon on Medium
Spring AI

Navigating AI for Developers with a Java Background

A practical guide for Java developers stepping into the AI world — covering frameworks, integration patterns, and how to leverage your existing skills.

Published on Medium
Engineering

From Monolith to Voice Agents: My AI Engineering Journey

Reflections on 11 years of software engineering and the pivot to building AI-native systems in production.

Coming soon on Medium

Skills & Technologies

12 years of accumulated tools, languages, and platforms.

💻
Languages
Java 17+PythonGoKotlinNode.jsJavaScriptReact
⚙
Frameworks
Spring BootSpring AIFastAPISpring SecurityHibernateFlaskGinJUnit
☁
Cloud & DevOps
AWSAzureKubernetesDockerTerraformJenkinsOpenTelemetryCI/CD
🗂
Data & Messaging
PostgreSQLMySQLRedisMongoDBCosmosDBKafkaRabbitMQAzure Service Bus
🤖
AI & ML
Agentic AILangChainRAGOpenAIClaudeMCPVector DBsPrompt EngineeringSpring AI
🛠
Tools
GitHubClaude CodeCopilotCursorIntelliJ IDEAVS Code

Let's Connect

Available for contract engagements. Let's talk about AI, engineering, or Manchester United.

Whether you need an agentic AI engineer on contract, have a collaboration idea, or just want to talk about the latest Old Trafford drama — drop me a line.