Lalit Surisetty.Software Engineer
WorkExperienceAboutContactGitHubResume
WorkExperienceAboutContactResume

All work

Five problems worth solving. Five ways I solved them.

01

doc-unify

Offline-first document intelligence pipeline & schema reconciliation engine.

FastAPIPythonTypeScript
explore→
02

Federated Learning, Live

Distributed client training pipeline streaming real-time convergence over SSE.

PythonFlowerFastAPI
explore→
03

Agentic Fraud Investigation System

XGBoost scoring microservice integrated with autonomous LLM multi-tool agent loops.

FastAPIPythonXGBoost
explore→
04

Oil Market Manipulation Analysis

Multi-engine computational pipeline evaluating time-series, causality, and RL policies.

PythonXGBoostARIMA
explore→
05

Temporal Drift in Accounting Concept Networks

Graph ETL pipeline & dynamic node2vec embedding alignment across quarterly SEC filings.

PythonGraph Embeddingsnode2vec
explore→

L.S.

Based in Hyderabad, IST (UTC+5:30) — happy to overlap with whatever time zone the work is actually in.

Reach

lalitsatvik@outlook.comGitHubLinkedIn

Open to ML, data, and backend work.

© 2026 Lalit Surisetty. Built with conviction, not a template.