One of 3 engineers building an AI-native CRM.
I primarily work on backend architecture, integrations, and data infrastructure powering the product.

• Integrations
  Built 10+ production integrations (Stripe, Salesforce, Linear, Google Calendar, and more).

• Analytics & Export Pipeline
  Rearchitected the analytics and export pipeline to reliably handle gigabyte-scale enterprise datasets, reducing exports from repeated timeouts to under a minute.

• Workflow Automation
  Built the workflow automation feature, enabling customers to define event-driven workflows using configurable triggers, conditions, and actions.

Built the backend architecture from scratch for an AI-powered equity research platform used by hedge fund managers on Wall Street.
The company raised $2.5M while the product was in beta.

• Infrastructure
  Designed and shipped the entire infrastructure solo: microservices, cloud deployment, data pipelines, and storage.

• LLM Pipeline
  Built an LLM pipeline that extracts and summarizes SEC filings and earnings calls at scale, cross-validating multiple models to reduce hallucinations before results reached analysts.

Early on, I optimized for shipping quickly because that's what the company needed. As we onboarded larger enterprise customers, some of those systems reached limits they were never originally built for.

I've since rearchitected many of them:

• Analytics queries serving millions of records with sub-second performance
• Export pipelines processing gigabyte-scale datasets reliably
• Large-scale data retrieval and filtering that remains responsive under enterprise workloads

Those rewrites taught me an important lesson: optimize for today's constraints, but always know which parts will eventually need to scale. I now design systems with those trade-offs in mind.

I think AI today is where computers were in the early days: powerful, centralized, and accessible to only a handful of organizations. Over time, computers became personal. I believe AI will follow the same trajectory, from cloud-first models to AI that is increasingly personalized, context-aware, and runs closer to the user.

As that shift happens, data becomes the real differentiator. The hard problems won't just be building models. They'll be collecting, moving, processing, and serving data reliably at scale while preserving privacy, security, and low latency.

That's why I've been investing heavily in distributed systems, data infrastructure, and backend engineering. Better models will matter, but so will the systems that move, process, and serve data at scale.

I document my engineering journey, the systems I've built, and my technical writing on my portfolio.

🌐 Portfolio: https://devaansh-bhandari.vercel.app/