Skip to content

Welcome to
Teja's Latent space.

An ai / machine learning engineer designing end-to-end intelligent systems — multi-agent workflows, RAG pipelines, and production ML infrastructure.

Get in touch
Impact

Numbers that shipped.

Outcomes from systems I've led or architected — across healthcare, enterprise, and data platforms.

0%

Operational overhead reduced

Multi-agent LangGraph automation for clinical reasoning workflows.

0%

Data processing efficiency

RAG pipelines on Snowflake, grounded with dimensional data models.

0%

Factual consistency

LLM quality bar enforced via RAGAS + G-Eval evaluation.

0%

Chatbot engagement uplift

Agentic enterprise assistant (Zuri) during trial rollout.

About

I turn data into autonomous action.

Strategic AI/ML engineer designing end-to-end intelligent systems. I lead multi-agent and RAG programs with Python, Advanced SQL, and distributed systems — with grit, adaptability, and a strong bias for shipping.

Agentic Systems

Multi-agent orchestration with LangGraph — closing the loop from raw data to autonomous action.

RAG & Big Data

Production RAG over Snowflake and dimensional data models, with hybrid retrieval and grounded outputs.

Evals & Guardrails

A "Quality Bar" for LLMs — RAGAS, G-Eval, and HITL loops for safety in high-trust domains.

Read my full story
Stack

Tools I reach for.

A pragmatic toolkit spanning research, engineering, and delivery.

PythonRustJavaC / C++Advanced SQLPyTorchTensorFlowTransformersCNNs / RNNsGANsLangChainLangGraphHugging FaceOpenAI APIsAWS BedrockAutogenCrewAIGoogle Gen AI SDKRAG PipelinesRAGAS / G-EvalLinear / Logistic RegressionDecision TreeRandom ForestPostgreSQLMySQLMongoDBDynamoDBSnowflakeAWSGCPAzureMPP / Distributed SystemsPythonRustJavaC / C++Advanced SQLPyTorchTensorFlowTransformersCNNs / RNNsGANsLangChainLangGraphHugging FaceOpenAI APIsAWS BedrockAutogenCrewAIGoogle Gen AI SDKRAG PipelinesRAGAS / G-EvalLinear / Logistic RegressionDecision TreeRandom ForestPostgreSQLMySQLMongoDBDynamoDBSnowflakeAWSGCPAzureMPP / Distributed Systems
Work

Selected projects.

End-to-end agentic AI, RAG systems, evaluations, and LM research — shipped.

Indian Enterprise Software Firm

ATLAS — Agentic Recruitment Intelligence

A bulk-resume ranker that reads past the buzzwords — grades candidates on actual work done and flags AI-sugarcoated bullets.

  • In production for months as the organisation-wide hiring system of record.
  • Handles 1,000+ resumes per role without a human bottleneck.
PythonLangGraphLangChainGemini
Enterprise Workforce Platform

Zuri — Unified Agentic Workspace

One conversation that replaces Jira, GitHub, email and HR tools — context-aware by role, privacy-respecting by design.

  • Collapsed 8+ daily tools into one conversation surface.
  • Role-aware views — IC, manager, leadership — without cross-leakage of PII.
PythonLangGraphLangChainOpenAI
Internal Product · Commercial Track

Qastra — Autonomous QA Agent Suite

End-to-end agentic testing — functional, regression, and live web journeys driven by browser-native agents.

  • Replaces brittle selector-based test suites with intent-driven test agents.
  • Autonomous regression loops triggered per deploy.
PythonLangChainLangGraphbrowser-use
US-based Healthcare Technology Client

Helix — Pervasive → Cloud AI Migration

Rescued a 6-month-stuck legacy-migration engagement — redelivered in 4 weeks with an AI-automated translation pipeline.

  • Closed a 6-month-stuck migration in 4 weeks.
  • Custom evaluation harness caught silent hallucinations the base LLM missed.
PythonLlama 3LangChainCustom Eval Framework

Have an idea worth building?

I'm currently open to selective collaborations and full-time roles. If you're building something ambitious with AI, let's talk.