Software Engineer - AI

Sundial Overview

Deep Comprehensible Insights from any Data

• Sundial is a top VC-backed early-stage startup headquartered at San Francisco Bay Area, US with a second office in Bengaluru, India. Our founders are industry veterans Chandra Narayanan, previously Chief Data Scientist at Sequoia Capital, and Julie Zhuo, previously VP, Design and Research at Facebook, author of a bestselling management book.
• We are a small team of top talent, high caliber Engineers, Data Scientists, Designers and PMs (currently 46 in India, 8 in US) and rapidly growing.
• We are on a mission to help builders make meaningful use of data to fulfil their vision. Sundial automatically diagnoses a product's data to explain the "what" and the "why" to enable faster and better decision making.

Our Vision for Sundial

• By now, you've probably visited our website and browsed around the profiles of our team members. If you haven't yet, please take a moment to do so. We'll wait 🙂
• Now you would probably know that we are focused on data storytelling. The data space has over $100B in market opportunity ahead of it, and modern Business Intelligence tools are growing at over 15% year over year.
• We've seen this evolution firsthand. Our co-founders Chandra Narayanan and Julie Zhuo cut their teeth scaling Facebook from a few million college students to billions of people.
• To make the best decisions possible, companies are investing more and more into understanding their data. And yet, demand far outpaces supply for Data Scientists and Data Platform Engineers who can construct useful narratives out of the growing firehose of raw data, tables and charts. Currently most data-centric organizations have a large Data Platform and Data Science team that builds Big Data Platforms, Insights Data stores to bring data into dashboards and manually generates reports to communicate the product story broadly. But a large part of this process can be easily productised.
• At Sundial, we're building a Sundial’s Insights Data Platform. This platform converts raw data in large Data Warehouses into a universe of deep product insights that product teams—including PMs, data scientists, executives and engineers—consume easily. This involves Highly Scalable, Robust Distributed Data Platform which can consistently, repeatably run complex Data Science and Transformation algorithms at Cloud Scale.

We envision a future where every organization becomes a data-informed organization through our work of:

1. Productising the diagnostic analysis of yesterday so teams can focus on the strategic bets for tomorrow.
2. Making data understanding easy and accessible to everyone, not just data scientists.
3. Surfacing opportunities of improvement in growth across segments

We believe better usage of data leads to better products, and better products lead to better experiences for people.

Why You’ll Love Working with Us

You like the ownership, camaraderie and chaos of a start-up environment

• Work at start-ups is very dynamic. Things move quickly and change frequently. We must be scrappy and flexible. Everyone will wear lots of hats.
• If you have future aspirations of being an entrepreneur or leader, you'll find few better learning grounds. You'll be given a ton of trust and responsibility. You'll see very transparently how we operate and make decisions. Your work will absolutely matter to the success of our company.
• Start-ups haven't "made it" yet. We have to convince customers we are valuable enough to them. You'll learn by doing and seeing how to operate in such ambiguous environments.

You love to build complex engineering systems at the intersection of data science and big data problems in the booming data analytics space.

• Data scientists unlock the secrets of data using complex algorithms. We believe that how they are presented are as important as the quality of secrets themselves.
• This poses novel and unique engineering challenges that require a creative and inquisitive mind for solving them. Some of the challenges are building a platform that enables many different ways to explore and explain the data.

You value learning and have a growth mindset

• Sundial is founded on the idea that slope is far more important than intercept. We are a learning environment, and all of us have something to teach and learn from each other. We invest heavily in learning sessions, sharing insights, and reflecting on our growth.
• We are also very well connected within the open-source frontend developer community (React, next.js etc) and lean heavily into that network to learn and make good decisions. Wondering what went into designing d3? Let's get on a call with Mike Bostock. ;)

You're interested in understanding how companies grow, and how data plays a role

• Unlocking the secrets of data is our bread and butter. How do successful companies grow? How do different types of businesses create value for users in an economically scalable way?
• If you find this area to be as fascinating as we do, that's awesome, because you're going to become an expert in this domain. :)

Sundial Insights Platform Overview -

warehouse-native SaaS data platform on AWS

Sundial has 5 primary Layers

Data Ingestion and ETL Layer

• This layer is a low-code/no-code ETL layer built on Spark. This can pull and transform Petabytes of data from Cloud Datawarehouses / Datalakes to Generate Metrics, Entity and Event Tables
• This layer typically processes 100s of Terrabytes of data daily to generate structured Sundial Data Models.

Sundial Datalake Store - Events, Entity and Metric Stores

• These are Datalakes which house structured Metrics, Entity Tables and Events.
• Sundial’s Insights Algorithms and Visualisations run on top of this Datalake Store.
• This mulit-tenant store is designed to scale to 100s of TBs of storage.

Sundial’s Insights Layer

• Sundial’s complex Data Science Algorithms are executed in this layer.
• This highly performant data processing layer is designed to execute complex Data Science Models to generate Insights.

Sundial’s AI layer

• Framework to build Generative AI Agents / Tools.
• These are used to serve complex Generative AI use-cases involving multi-step / dynamic AI Agent use cases like Insights Report Generation, Data Modelling and Ad-Hoc Insights Exploration

Sundial’s Visualisation Layer

• Representing Insights are easily comprehensible and explorable for anyone.
• The Visualisation layer is rich and complex . With latest technologies powering a seamless experience.

The technologies which power the above systems are :

1. AWS Services like EKS, EMR Spark, S3, RDS, Athena and Opensearch
2. Our stack is built using frameworks in Go, Python, Pyspark, Pandas, Typescript, Terraform etc.
3. We use advanced Devops systems and practices for the best developer productivity.
4. We deal with Big Data Processing systems and OLAP systems which has to process Petabytes of data for some of the fastest growing companies.
5. Frontend is built using Typescript, React, Redux and GraphQL
6. Quality is something we care about a lot. We expect every module to maintain a very high bar through all best practices related to Code Quality and Test Driven Development.

As an AI Engineer, are you excited about…

• Building a scalable AI agent:

  We are releasing a state-of-the-art AI agent based on proprietary knowledge, rather than using a cookie-cutter agentic framework. Doing so requires innovative thinking, and the ability to deal with probabilistic development.

• Writing Exceptional Prompts:

  You have spent hours honing techniques for writing clear instructions and picking good exemplars to feed into an LLM. You have good intuition about the limits of what a model can handle.

• Building Systems for Scale:

  Design and build ML-Ops pipelines to monitor model performance and data drift. Provide guidance on optimizing the process for scale.

• Preaching an AI-native Mindset:

  You are an evangelist for thinking about building products in an AI-first manner, which means you are comfortable with the non-deterministic nature of AI outputs.

• Cross-Functional Collaboration:

  Clearly explain technical concepts to a diverse range of audiences spanning engineering, product, data, and execs to deliver AI solutions aligned with user needs.

Are you…

• Seasoned AI Engineer with 2+ years of hands-on industry experience applying AI or machine learning in production systems (LLM, data, or systems contexts)
• Hold an Undergraduate or Masters degree in Computer Science or related field. Equivalent work experience will also suffice; we are more concerned with your ability to think rather than your pedigree.
• Have strong programming skills in Python, and familiarity with modern AI best practices around prompting, context engineering, and agentic evals.
• Experience with cloud APIs, cloud infrastructure (AWS, GCP, or Azure), and enterprise integrations
• Understanding of basic deep learning concepts such as gradient descent, how to prevent overfitting, and precision vs. recall trade-offs.
• Able to prioritize against competing constraints. Great product development requires striking the right balance between moving quickly while minimizing technical debt.