// not a portfolio — an agent-driven profile of one engineer

Hi, I'm Suresh.
I build agentic systems, full-stack apps, and data at scale.

Founding Engineer at Firmable, first engineering hire turned architect of its AI data stack. Fifteen years shipping agents, LLM pipelines, full-stack products, and data platforms from the very first byte to billions of records and trillions of tokens, and building the teams behind them.

agent — grounded on profile.yaml
suresh@agent:~$
agent ▸ trace(career.pipeline)

Career, rendered as a DAG

Nov 2022 → present · full-time · remote

Founding Engineer RUNNING

Firmable · AI-powered sales intelligence

First hire; architected the AI data stack and built the team.

2020 → 2023 · three tasks, one scheduler

The parallel years ∥ CONCURRENT

Data Fusion Labs · Import.io · Accord Property Services
  • CEO, Data Fusion Labs (2020-23): ran a news-as-a-service platform for enterprise clients
  • Enterprise Solutions Architect, Import.io (2020-22): data architectures at 2 trillion+ records monthly
  • Back End Developer, Accord (2022-23): secure, scalable AWS backends for a construction-tech platform
Apr 2018 → Jan 2023 · 4 yrs 10 mos

Principal Software Engineer SUCCESS

Dell Technologies

Architected cloud-native microservices and distributed data systems.

Jul 2011 → Apr 2018 · 6 yrs 10 mos

Technical Lead ← Programmer Analyst ← Trainee SUCCESS

Cognizant

Built crawlers powering job listings in Google's jobs experience.

agent ▸ fetch(pretraining)

Pretraining

degree
B.Tech, Computer Science
school
Sreenidhi Institute of Science and Technology
years
2008-2011

// where the base model was trained.

agent ▸ recall(operating.principles)

How I operate

get_it_doneGet the work done — the tool is an implementation detail. Python crawlers, AI agents, or a brand-new team: whatever ships the outcome.
zero_to_oneBuild from scratch. Startups, teams, and data stacks — from the first byte and the first hire.
evals_firstNever run an LLM pipeline blind. Evals before scale, always.
ai_nativeIf an agent can do it, an agent should. Humans do the judgment.
cost_per_inputUnit economics are first-class. Token-max the useful, cut the slop.
direct_commsTerse beats verbose. Ship fast, measure everything, say the thing.
agent ▸ fleet.list()

The fleet

Specialist agents, each defined by its own skill file. Activate one to interrogate it.

agent ▸ interrogate(fleet)

Interrogate the fleet

agent — pick a specialist
# pick an agent, then a question. answers are baked into each skill file.
agent ▸ schedule(call)

Book a slot

Data platform architecture, LLM pipelines, AI-native teams — or just a good conversation. No email on this site; the calendar is the API.

calendly.com/sureshbpro ▸ DM on LinkedIn