About

Lead Applied AI Engineer building multi-agent architectures, agent harnesses, and LangChain-based AI systems

The throughline is consistent: build AI systems that are grounded, production-minded, and understandable to the teams who need to operate them.

Professional bio

I’m a Lead Applied AI Engineer with over 7 years of experience across NLP, Generative AI, and production ML systems. My core work today is around multi-agent architectures, agent harnesses, LangChain-based systems, and retrieval-augmented generation.

I like working at the layer where model capability, system design, and product usability meet. That usually means building workflows that are not just impressive in isolation, but grounded enough for operations teams, analysts, and business users to actually depend on.

A lot of my background also comes from finance and enterprise document-heavy environments, which shaped how I think about reliability: explicit orchestration, measurable behavior, grounded outputs, and systems that teams can inspect and evolve after launch.

Career path

The path from applied NLP to production GenAI systems

01

2023 to now

Manager AI/ML Expert, Generative AI

Genpact

I design and ship production GenAI systems, including LangChain-based RAG products, privacy-first local LLM deployments, documentation assistants, and CXO-facing analytics experiences.

02

2022 to 2023

Lead MTS II

Alphastream Technologies

I led NLP and Python developers across fixed-income workflows, fine-tuned and deployed open-source LLMs, built end-to-end fund data pipelines, and contributed to microservice architecture for financial products.

03

2019 to 2022

Junior Data Engineer

Almug Technologies

I built and deployed NLP systems for document extraction in financial environments, owned feedback-driven modules used by analysts, and introduced stronger Git and test workflows to improve software stability.

04

2019

Data Science Intern

Attune Technologies

I worked on chest X-ray nodule detection and image-processing pipelines, which gave me an early foundation in applied ML before I moved fully into NLP and GenAI systems.

Primary focus

Work aligned to current AI product and platform needs

01

Multi-Agent Architectures

Planner, orchestrator, and specialist-agent systems with explicit state and human approval loops.

02

Agent Harnesses

Evaluation harnesses, retries, observability, tool safety, and sandboxed execution.

03

LangChain + RAG

LangChain-based retrieval workflows, grounding, and production-ready knowledge systems.

Skills and expertise

Architecture, orchestration, and applied LLM delivery

Multi-agent architecture design and orchestration patterns

LangChain workflows for tool use, retrieval, and stateful agents

LLM application design for enterprise and analytics teams

Human-in-the-loop patterns for safety and iteration

Agent harnesses, evaluation loops, and runtime guardrails

Retrieval architecture and grounded generation

Observability-minded system design for AI products

Working principles

How the work is evaluated

01

Design for reliability before demo polish.

02

Keep agents transparent, inspectable, and measurable.

03

Ship architecture that teams can maintain after launch.

04

Treat UX as part of system quality, not decoration.

Outside work

I also like building things beyond my core day job

Outside my main AI engineering work, I enjoy shipping products end to end. That includes building and publishing a game on the Play Store and creating `krutrima.in`.

Those side projects matter to me because they keep my thinking broad. They force me to care not only about models and backend systems, but also about product decisions, user experience, deployment, and what it takes to turn an idea into something people can actually use.