Deeksha Gautam
Deeksha Gautam

Status — Available • Focus — Applied AI • Based — Pune, India

I create AI agents
that solve real-world problems.

I'm Deeksha Gautam, a software engineer with four years of experience
building enterprise applications and AI-integrated systems. I design dependable
LLM workflows with guardrails, structured outputs, evaluation, and fail-safe handling—then connect them to backend services and production delivery.

View projects →Download résumé

Experience

Where I'm building now.

Primary role

2022 — Present

Technology Analyst · Citi

Building and operating enterprise platforms across AI automation, backend services, event-driven systems, and production delivery.

  • Built a three-level NAM fund-discount approval workflow integrated with Salesforce, reducing turnaround time by 78%.
  • Created an AI agent for production change management and documentation, reducing human intervention by 75%.
  • Developed a Kafka-powered investment platform that contributed to 47% client growth and 30% higher activation.
  • Migrated and deployed 35 Spring Boot services across APAC, NAM, and EMEA with no production issues; automated workflows reduced manual effort by 65%.

Advisory side project

Sep 2026 — Present

Technical AI Advisor · CareAmy

Designing CareAmy's migration from externally hosted LLM APIs to a fully in-house, privacy-first patient-support chatbot. Building a production-grade RAG pipeline spanning document ingestion, embeddings, hybrid retrieval, context orchestration, controlled responses, guardrails, and evaluation. Work in progress.

Current focus

LLM systems

Turning language models into task-focused agents, automation, and usable applications.

Evaluation & safety

Building guardrails, structured outputs, fallbacks, and automated evaluation into each workflow.

Production engineering

Connecting AI to APIs, backend services, and CI/CD with enterprise delivery discipline.

Featured work

All projects →
01Hybrid AI Gateway

Privacy-aware model routing with 106 automated tests.

→02Local vs Cloud AI Lab

A practical benchmark comparing Gemma 3 4B running locally through Ollama with GPT-OSS 120B running through GroqCloud.

→03Local RAG Evaluation Lab

A privacy-first local RAG system that combines dense vector retrieval and BM25 through Reciprocal Rank Fusion, then evaluates retrieval quality before generation.

→

Side quest

Writings

Medium profile ↗
Sep 13, 2024

Amazon’s AI Evolution: The Future of Alexa and Beyond

Jul 31, 2024

Apple’s AI Secret: What’s Behind the Google Partnership?

Jul 17, 2024

Web 3.0 : Revolution or a Rip-off?

Jun 24, 2024

Neuralink: Is It Truly the Pioneer?

May 31, 2024

Ambani’s Twist on Quick Commerce

May 18, 2024

While Google I/O 2024 unveiled groundbreaking tech, Many of these advancements remain out of reach of many Indian users

Interested in dependable LLM systems or enterprise AI automation?

Let's talk ↗