Skip to content

I'm Sreevarshan, a Boston-based engineer building intelligent systems, data platforms and software products.

NowFinishing my M.S. in Data Analytics Engineering at Northeastern

01 · What I do

Pipeline, model and product.

I work across all three, because an AI feature is only as good as the data it reads and the interface that explains it. Most recently I built Snowflake Cortex AI agents and MCP integrations at RightRev.

02 · What I'm working toward

AI that teams can trust.

Agents that cite their sources, data that is checked before it is used, and results I can measure.

Focus areas

  • Agentic AI systems with verifiable outputs
  • Retrieval, evaluation and LLM reliability
  • Streaming and batch data pipelines
  • Full-stack products on top of data platforms

A.1Values

How I work.

01

Evidence over adjectives

Every number I report comes from a test, an eval run or a log. If I can't verify it, I say so.

02

Guardrails are features

Tenant isolation, RBAC, input validation and abstention belong in the first version, not the third.

03

Data is the product

Schema checks, lineage and quality gates decide whether anything downstream can be trusted.

04

Explain the system

Interfaces should show their work: the source, the query, the reason a score was given.

A.2Capability map

Skills, with receipts.

No percentage bars. Select a skill to see the roles and projects where it was actually used.

Intelligent systems

Agents, retrieval and evaluation that hold up under verification.

Data engineering

Pipelines, models and quality gates that make downstream work trustworthy.

Full-stack software

Product surfaces and APIs that explain what the system did.

Security, quality & delivery

The parts that let a system be trusted in production.

Evidence

AI agents