sybill.ai

Sybill is an emotionally aware video call assistant for sales teams. Sign up for free at https://www.sybill.ai.

Sybill allows salespeople to understand their prospects more deeply with sentiment and engagement insights. Never take notes on a call again, and manage your deal with prospect intelligence. 

Sybill allows sales leaders to transform their messaging and decks based on what resonates with their prospects, and coach their teams on winning behaviors. 

Our founding team comes from Stanford, Harvard, and UCSD, and has a deep background in machine learning and human psychology. 🙋🏻‍♀️

Co-Founder and CTO

Forbes Technology Council

As an official member of this invite-only council for CTOs/CIOs, I will be contributing my ideas to my peers in the industry through panel discussions, and original articles on Forbes.com

Official Member

Confluent

At Confluent, my primary role as a platform engineer was to design and build the tiered storage layer that made Apache Kafka much more elastic and capable to serve as the system of record for the Fortune 100 companies.

Platform Engineer

University of California San Diego

Graduate Teaching Assistant

Salesforce

At Salesforce, my role primarily involved designing and evaluating the open-source Envoy proxy in a service-mesh architecture through statistical load (latency + throughput) analysis using Docker and Kubernetes for deployment purposes and Go as the programming language.

Software Engineer Intern

Morgan Stanley

Designed and developed a distributed Bitemporal data access layer in Functional Scala. This project was recognized in the Circle of Excellence within the Firm.

Senior Associate Technology

Technology Analyst Intern

Arista Networks

Software Engineer Internship

Robotics and Artificial Intelligence Laboratory, IIIT Allahabad

Summer Intern

Identify what resonates with your buyers with Behavior AI | Co-founder at Sybill - Hiring across technology roles

Hi, welcome to my profile! I'm Soumyarka, Co-Founder of Sybill, and a UC San Diego and IIIT Allahabad alumnus.

Check out Sybill ---> Sign up for free at https://www.sybill.ai for your emotionally aware call assistant.

Want to chat with me? ---> https://calendly.com/soumyarka/meet. I love chatting about sales tech, understanding humans better, AI and startups. 

What is Sybill? ---> It’s an emotionally aware video call partner, leveraging body language understanding and AI to help sales teams understand their prospects better, accelerate revenue, and improve themselves and their collateral.

I graduated from UC San Diego with MS in Computer Science (AI) with research experience in the intersection of Distributed Systems and applied Machine Learning.

Prior experiences include software infrastructure teams within Confluent and Salesforce in the San Francisco Bay Area and Morgan Stanley in Mumbai.

Taking an AI system to 10k users = hard design choices.

Here are some pitfalls I hope you don't repeat:

1. Multi-agent coordination hell

Talked to teams who spent weeks debugging inconsistent answers.
The issue? Multiple agents. Different contexts. Independent reasoning. Contradictions.

The "solution"? Another agent to reconcile.

You see the problem.

2. Semantic search breaks fast

"Why is the Acme deal stalled?"

Vector DB finds documents about "stalled deals."

What you need: Sarah = S. Johnson. Features across 6 calls. Objections in Slack.

Semantic similarity ≠ structured reasoning.

~~ The real trap ~~

Most sales AI looks great in demos.

Six months later? Unmaintained coordination. Inconsistent graph quality. Model upgrades break everything.

Architecture matters more than the model.

I wrote about the journey from semantic search to structured RAG. The trade-offs that aren't obvious until production.

I have posed some architectural questions that I would love to learn from each of you.

Mehak Aman Prajwal Nishit Gorish Sybill

#AIArchitecture #Learnings #AgenticReasoning #SalesAI

Soumyarka Mondal

Cofounder & CTO at Sybill.AI - Hiring across engineering, product & ML roles

Part 1 of 3: What We've Learned About Single vs Multi-Agent Architectures and Structured Retrieval After building Sybill and working with hundreds of 

Open article: Wins and pitfalls from designing an AI System used by 10k+ users by Soumyarka Mondal • 4 min read

Wins and pitfalls from designing an AI System used by 10k+ users

taplio

Soumyarka Mondal's Linkedin Analytics

Soumyarka Mondal

Performance ScoreSoumyarka Mondal's LinkedIn Performance

Soumyarka Mondal's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Soumyarka Mondal's Linkedin Analytics

Soumyarka Mondal

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Soumyarka Mondal's LinkedIn Performance

Soumyarka Mondal's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Soumyarka Mondal's Linkedin Analytics

Performance ScoreSoumyarka Mondal's LinkedIn Performance

Soumyarka Mondal's Best Posts (last 30 days)