01
AI agent systems
Turn ambiguous, human-heavy workflows into tool-using systems with grounding, controls, evaluation, and human review.
AgentsRAGEvals
Based in Hyderabad, IndiaAI agents · Automation · Data
AI Agent & Automation Engineer
AI systems
that do
real work.
I build AI agents and automation systems that turn complex financial workflows into reliable infrastructure.
Observe → Reason → Act → VerifyJudgment + deterministic controls
AI orchestrator
Routing → Planner
AI orchestrator
Routing → Planner
Planner activeBreaking the objective into verified tasks
01 / Hire me
Hire me for
Bring me an operational workflow that is slow, brittle, or too dependent on manual judgment. I can scope the system, build the reliable path, and leave it observable.
Hyderabad, India · remote-readyConsulting · partnerships · rolesAI agents · automation · data
Start a conversation 02
Automation engineering
Replace repetitive browser, document, reporting, and operations work with recoverable, production-ready automation.
PythonPlaywrightWorkflows
03
Data & intelligence platforms
Build dependable ingestion, transformation, quality, and decision layers for financial and operational data.
GCPBigQueryOrchestration
Measured impact
Proof over possibility.
84%
Editorial agreement
AI financial-news filtering measured against human editorial decisions
8
Production systems
Shipped across data, research, compliance, onboarding, and reporting
~1,500
Hours saved annually
Estimated recurring manual effort removed across operational workflows
System method
From messy input to accountable output.
AI is one component in a reliable system—not the entire system. The surrounding architecture makes its work observable, reviewable, and safe to operate.
Observe
Ingest market data, research inputs, documents, and human intent with traceable source context.
Reason
Apply financial-domain rules and model judgment where ambiguity needs more than a conventional script.
Act
Call tools, transform data, create reports, and move work through operational systems.
Verify
Validate schemas, evidence, outputs, and failure paths before the system hands work back to people.
Stack27 technologies
Tools I build production systems with.
Selected / LangChain
Composable LLM applications and tool-connected workflows.
01AI & agents
02Application
03Data & cloud
04Delivery
Selected systems
Built around the workflow.
View all01 / AI editorial agent
Financial News Intelligence
A production pipeline that turns a noisy overnight market-news stream into a focused, reviewable editorial queue.
4–5 hours → 8 minutes2 overnight operators → 084% editorial agreement
Explore case study Editorial agent trace84% agreement
02 / Document automation
Research Report Factory
A controlled reporting system that turns structured stock research into validated, chart-rich client documents overnight.
120–200 DOCX reports5–7 days → overnightAutomated validation gates
Explore case study Report factory200 reports
03 / Execution workflow
Broker Order Orchestration
A structured relay that converts investment instructions into broker-ready order artifacts with validation and auditability.
Instruction → CSV in ~10 sec21-column schemaDesigned for multi-client operations
Explore case study Order compiler21 columns
04 / Data infrastructure
GCP Market Data Platform
A medallion-style financial data platform for reproducible ingestion, transformation, quality checks, and analytics.
Bronze / Silver / GoldInfrastructure as codeAnalytics-ready models
Explore case study Medallion platform3 governed layers
01
Agent systems
Tool-using workflows, grounded reasoning, structured outputs, human review, and evaluation loops.
02
Automation engineering
Reliable browser, document, email, and operational automations designed around failure and recovery.
03
Data platforms
Cloud ingestion, transformation, quality contracts, orchestration, and analytics-ready data models.
04
Financial workflows
Applied knowledge of Indian markets, research operations, KYC, compliance, reporting, and broker formats.
Have a hard workflow?Let's make it operable.