Swiggy Tech Organization – Leadership roles and responsibilities

Objective: This document proposes a Tech Org structure for Swiggy, with description of roles and responsibilities of key
tech leadership roles. The intent of this document is to provide clarity to the organization by defining the tech org
structure, including role scope of tech leaders, to all internal stakeholders (i.e. Swiggy tech team, product team, business
teams, recruiting team, external recruiting agency).

Background: The Swiggy Tech team is a XXX size tech team with YYY BIS as of June 1st, 2018. The tech team is collocated
with the product and business team in the Swiggy office in Bangalore. The tech team currently reports into the CTO via
mid-level and line managers. The tech team is missing AVP/VP level leadership (one level removed from the CTO) and is
aggressively hiring to fill this gap.
In the current state, the tech team organization is fluid, with the team responding in an agile and flexible manner to
business needs. Corporate projects are driven through a “big bets” roadmap which defines a clear top-down priority list
for cross-company wide initiatives. The product team owns building the Product requirements (PR/FAQ and PRD) for
these projects. Once the PRD is available, the tech team self-organizes to staff and execute against these big bets. Once a
big bet is delivered, the team can be repurposed for other big bet projects. The Big Bets process has worked well to
mobilize the org to deliver strategic priorities. However, operational excellence investments have often been
deprioritized due to limited engineering bandwidth.
Looking ahead, Swiggy has an ambitious plan to grow the core (food delivery) business while also expanding Swiggy into
new verticals. The product team is scaling to meet this need by restructuring into 4 product verticals – (a) Storefront, (b)
Delivery, (c) ERP and (d) Revenue and Growth. As the complexity of our product and business lines increase, the Swiggy
Tech org will also have to scale to meet this challenge and become a key competitive advantage for the future.

Tech Organization guiding principles: The guiding principles behind the proposed Tech organization are –
a. Build business empathy and co-ownership of product and business goals – We will build a tech organization that is
very closely aligned with the business and product goals. The tech organization will co-own the business/product
goals and be an equal partner in the definition of the product roadmap. We will organize the tech team around our
different customer segments such (a) the end consumer, (b) our business Partners (a.k.a. vendor, supplier, restaurant
owner), (c) the delivery associate and (d) internal Ops teams such as CS Ops, delivery Ops.
b. Build long term-thinking and accountability via software ownership – Every tech team will own 1 or more software
components with clear boundaries of scope and impact. With ownership, teams will be responsible for charting the
technology roadmap by working backwards from the current and future business/product needs. In addition to
business / product goals, teams will also own technical fitness functions that measure the performance of their
software.
c. Speed matters – We are in a ultra-competitive market and speed of execution matters. While we lay the foundation
blocks of a scalable tech organization, we will continue to be flexible in organizing to staff and execute new and
emergent business needs. We will bias towards experimentation and light-weight implementations to execute fast.
Once we get early validation from our customers, we will double down on our technology investment.

Proposed Tech Organization structure:

Effective XX/2018, the Swiggy Tech organization will have both a CTO and a Head of Engineering.

The CTO office will own the following functions –

a. Swiggy Data Platform (i.e. Data Warehouse and data ingestion platform)
b. Info-security
c. IT systems
d. Developer Tools
e. Corporate Systems (HR, Performance Management, Central Accounting

The Head of Engineering will own product engineering and data science for Swiggy. This includes –
a. The Consumer Experience (VP, Tech) – This team owns the consumer journey from customer acquisition through
conversion or order creation. (i.e. the Swiggy App/website and underlying CX and services)
b. The Delivery Experience (VP, Tech) – This team owns the delivery experience, including the order promise, delivery
and post-order support. (i.e. Supply Chain Optimization, Order Fulfillment and Driver Experience)
c. The Partner Experience (AVP, Tech) – This teams owns the products and initiatives that help our business partners
succeed with Swiggy (i.e. Restaurant owner tools, Advertising, Promotions etc)
d. Swiggy Operator Experience (AVP, Tech) – This team owns the internal facing tools and products that help our
Business Ops teams scale and delivery a great customer experience. (i.e. Customer Service tools, Fraud
Investigations, Swiggy Ops tools, Biz Dev tools)
e. Data Science and Analytics (VP, Data Science) – This team owns improving the product experience and business
performance by applying statistical methods and AI models to our product (i.e. Analysts, Applied ML and Research)
1
Company Confidential Swiggy Tech Organization – Leadership Roles
Swiggy Engineering Organization – Detailed view

This section describes the role scope of the N-1 leaders under the Head of Engineering.

A. VP, Consumer Experience Tech

Role scope: Owns the engineering for the Swiggy customer journey from the time we acquire the customer, till the time
the customer places the order. This includes customer acquisition via paid or free traffic, product discovery and
personalization, customer identity, transaction and payment experience and customer engagement to drive repeat
usage. This team will co-own the goals for both Big Bets and Brilliant Basics that impact Consumer Experience.

System ownership: This role owns the Swiggy app, website and underlying services that power the consumer experience.
On the Experience layer, the team will own all the UI components (app and website) across all supported platforms
(Android, IOS and Browser). On the service layer, the team will own the Swiggy Catalog including the feed ingestion
pipeline, Data Aggregation APIs for the CX, Search, Personalization, Cart, Identity, Location Detection, Consumer
Payments and the Ordering platform.

What are some of the opportunities for innovation and tech leadership in this role?

- Building a multi-tenant, flexible, scalable and authoritative Swiggy Catalog – As we expand into new categories and
scale the business across more cities and countries, we will ingest data from varied sources, each at varying levels of
data quality and product authority. The Swiggy Catalog will power Consumer Experience across these businesses and
products. This team will own building a multi-tenant catalog system that is (a) flexible to ingest data across different
categories, (b) scalable to process data writes and low-latency data reads (c) intelligent (AI enabled) in item
normalization, matching and fact extraction to build an authoritative catalog.

- Personalized Search and Recommendation – As we learn more about our customers daily needs, we can provide
them with a personalized search and discovery experience that is tuned to their individual needs and preferences. For
example, (a) identify a customer’s stated or implicit preference or constraints based on their previous order history
and similar cohorts (e.g. identifying a customer’s preference for Jain food), (b) extract insights about our catalog based
on order history (e.g. identifying a restaurant with the highest rated Jain food dishes), (c) target customers with
personalized, timely and relevant recommendations based on these insights (e.g. Recommend a top Jain food
restaurant with a promo coupon, early in the evening).

- Intelligent Ordering (One-click ordering) – For our most engaged customers, we can simplify their life by analyzing
their consumption pattern and predicting future demand. For example, we can predict if a customer is likely to order
pizza on a certain day of the week. For this customer, we can also recommend a highly rated pizza, with a discount,
and give the customer the option to place the order with a single click.

- Voice and Chat based shopping (including local Indian languages) – Over the last few years, conversational search
(i.e. Voice and Chat) has become increasingly popular in India. Our customers will also want to interact with Swiggy
and discover and order services using Voice and/or Chat interfaces. We will need to invest in Natural Language
Processing, Voice Recognition and Machine Translation to serve our customers, even in local Indian languages.

2
Company Confidential Swiggy Tech Organization – Leadership Roles
B. VP, Delivery Experience Tech

Role scope: Owns the engineering for the Swiggy delivery experience, including the delivery promise, the delivery
experience and post-order care. This includes (a) Supply Chain Optimization including capacity management, real-time
serviceability, delivery promise, order batching and order assignment, (b) Fulfillment experience including store
fulfillment experience, order exception management, (c) Delivery experience including delivery associate onboarding,
training, delivery execution and tracking, delivery associate payments and performance management, (d) post-order care
such as order cancelation, item cancelation and refunds. This team will co-own the goals for both Big Bets and Brilliant
Basics that impact Delivery Experience.

System ownership: On the Experience layer, the team will own the Driver Android app and the consumer facing
components for Order Tracking and Post-order care on the app and website across all supported platforms (Android, IOS
and Browser). On the service layer, the team will own (a) Supply Chain Optimization systems for capacity planning,
demand shaping, serviceability calculation, delivery promise calculation, order batching, and order assignment (b) service
functions for Order and Item level Refunds and Cancelation and (c) service and data layer for Drivers and (d) Store Order
Fulfillment services

What are some of the opportunities for innovation and tech leadership in this role?

- Building a real-time on-demand supply chain engine, powered by algorithms and AI – At the heart of the Swiggy
delivery network is the supply chain optimization engine. The current supply chain network is powered by a mix of
business rules, operational processes and a few algorithms. In order to build a cost-effective and scalable delivery
network, we will have to invest in AI / ML based systems that will analyze thousands of data points in real-time and
make supply chain optimization decisions in milli-seconds. Some of the key optimization problems to solve include –
(a) Predicting the delivery capacity of our network in a zone, based on predicted order volume / restaurant load /
delivery associate availability etc. (b) real-time shaping of consumer demand to react to network issues such as
restaurant performance, weather, traffic congestion etc, (c) Identifying how and when to batch nearby orders, to
minimize delivery cost while constraining to customer promise, (d) improving delivery promise accuracy by predicting
restaurant prep time, delivery time etc and more.

- Swiggy Maps – Our Swiggy delivery associates travel XX kilometers in a day, traversing through nooks and corners of
Indian cities. While we currently use 3rd party sources for map data, our ambition is to leverage the on the ground data
from our delivery associates and build an optimized Swiggy Maps dataset. This dataset would have augmented data
such as parking area for individual buildings / restaurants, real-time road closures, waypoints for order consolidation
etc. The Swiggy Maps dataset would be used to improve our on-time delivery by improving accuracy of promise as
well as guiding delivery associates out on delivery.

- Building a highly trained and engaged fleet of delivery associates – Swiggy has one of India’s largest delivery fleet of
XX thousand delivery associates. As we scale, we will multiply our delivery fleet. Our delivery fleet is critical to the
Swiggy product promise and we need to provide Delivery Associates with the right training, incentives and support to
make them successful on Swiggy. This team will have to build a vernacular (local language optimized) experience with
innovative products such as Driver Gamification for learning, Incentive / Loyalty programs, real-time interventions,
driver performance management etc.

- Fulfillment Technology – A key part of the fulfillment experience is the time taken by the Node (e.g. Store /
Restaurant) to fulfill or prep the order. We will optimize fulfillment of the order by investing in technology solutions
such as Store/Restaurant POS integration, mobile solutions for operators, warehouse management systems etc.

3
Company Confidential Swiggy Tech Organization – Leadership Roles
C. VP, Data Sciences
Role scope: This role owns improving the product experience and business performance by applying statistical methods
and AI models to all relevant facets of the product. The role will own (a) Statistical Analysis for building heuristics based
business rules, (b) building ML models for predictive and recommendation systems, (c) research into long term bets that
require sustained focus and (d) being a force multiplier by training engineers in applying off-the-shelf ML models.

System ownership: The Data Science team will build algorithms and implement them in scripts using either R or Python.
The team will leverage the Data Platform (built by the CTO office) for their data needs. They will also work with the
engineering teams to productionize the ML models.

What are some of the opportunities for innovation and tech leadership in this role?

This team will own making Swiggy an AI-first company, by applying AI / ML to all aspects of the product. Some of the
problems to solve include:

- Building a real-time on-demand supply chain engine, powered by algorithms and AI – At the heart of the Swiggy
delivery network is the supply chain optimization engine. The current supply chain network is powered by a mix of
business rules, operational processes and a few algorithms. In order to build a cost-effective and scalable delivery
network, we will have to invest in AI / ML based systems that will analyze thousands of data points in real-time and
make supply chain optimization decisions in milli-seconds. Some of the key optimization problems to solve include –
(a) Predicting the delivery capacity of our network in a zone, based on predicted order volume / restaurant load /
delivery associate availability etc. (b) real-time shaping of consumer demand to react to network issues such as
restaurant performance, weather, traffic congestion etc, (c) Identifying how and when to batch nearby orders, to
minimize delivery cost while constraining to customer promise, (d) improving delivery promise accuracy by predicting
restaurant prep time, delivery time etc and more.

- Personalized Search and Recommendation – As we learn more about our customers daily needs, we can provide
them with a personalized search and discovery experience that is tuned to their individual needs and preferences. For
example, (a) identify a customer’s stated or implicit preference or constraints based on their previous order history
and similar cohorts (e.g. identifying a customer’s preference for Jain food), (b) extract insights about our catalog based
on order history (e.g. identifying a restaurant with the highest rated Jain food dishes), (c) target customers with
personalized, timely and relevant recommendations based on these insights (e.g. Recommend a top Jain food
restaurant with a promo coupon, early in the evening).

- Voice and Chat based shopping (including local Indian languages) – Over the last few years, conversational search
(i.e. Voice and Chat) has become increasingly popular in India. Our customers will also want to interact with Swiggy
and discover and order services using Voice and/or Chat interfaces. We will need to invest in Natural Language
Processing, Voice Recognition and Machine Translation to serve our customers, even in local Indian languages.

- Building an authoritative Swiggy Catalog – As we expand into new categories and scale the business across more
cities and countries, we will ingest data from varied sources, each at varying levels of data quality and product
authority. The Swiggy Catalog will power Consumer Experience across these businesses and products. This team will
own building ML models for item normalization, fact extraction and matching to build an authoritative catalog. For
example, we will need to use Deep Learning to analyze product images and extract facts for assessing data quality.

- Abuse detection and prevention – An important function for the Data Science team will be to build abuse detection
models that minimize abuse from customers, delivery associates and/or partners. Some example fraud detection
models include duplicate customer account detection (created by abusive customers to avail new account
promotions), customer abuse score (to predict propensity of a customer to claim a fraudulent refund), delivery
associate abuse score (to detect propensity of a delivery associate to cancel a delivery) etc.

4
Company Confidential Swiggy Tech Organization – Leadership Roles