Production AI systems

Krishna builds reliable AI products from data to deployment.

Machine Learning Engineer with 3+ years of experience building multi-agent assistants, forecasting platforms, RAG applications, data pipelines, and cloud-native services that teams can trust in production.

Start a conversation View GitHub LinkedIn
Agent workflows Assistants that plan, answer, retry, and report progress.
Model platforms Serving, tracking, deployment, observability, and cost control.
Data products Revenue forecasting, EDA, feature design, and validation gates.
Cloud delivery Docker, Kubernetes, EKS, ArgoCD, CI/CD, and stable rollouts.

At a glance

Recent work with measurable business impact.

Recent systems have reduced manual review, lowered LLM usage costs, accelerated long-running workflows, and improved visibility for enterprise teams.

60% less manual review during ERP requirement discovery through an AI workflow assistant.
40% lower LLM API cost with prompt-aware caching, output reuse, and template testing.
25% faster multi-stage processing using queued jobs, background workers, and retry handling.
135 public GitHub repositories, with recent work across RAG, guardrails, and deep learning.

Experience

Building AI, data, and cloud systems at Artha Solutions.

Current role: Software Development Engineer in Scottsdale, Arizona, focused on production ML workflows, data platforms, and enterprise delivery.

AI-Powered Workflow Assistant

Multi-agent system for ERP discovery, planning, task tracking, and project status answers.

  • Built agents that analyze organizational knowledge bases and reduce time spent reviewing documents.
  • Added caching and versioning for LLM outputs so teams can reuse answers and compare prompt templates.
  • Designed queued background jobs with live progress tracking for long-running AI workflows.
  • Deployed Dockerized microservices on EKS with ArgoCD-managed releases and stable rollouts.
LLM workflows PostgreSQL Docker EKS ArgoCD Microservices

Revenue Forecasting Platform

Explainable forecasting for regional revenue planning under changing business conditions.

  • Explored ten years of revenue data with PySpark, Hive, Impala, and PostgreSQL.
  • Designed time-series features including lags, rolling measures, volatility, regional patterns, and macroeconomic inputs.
  • Mapped a path from analysis to production with MLflow, validation checks, drift monitoring, and retraining.
Forecasting PySpark Hive Impala MLflow PostgreSQL

ThinkTank Lab

Inventory forecasting prototypes for Fortune 500 stakeholder decisions.

  • Led three interns through prototype delivery, model evaluation, and stakeholder-ready findings.
  • Connected Hive data through Impala and moved results into Azure Synapse for analysis.
  • Used Python and H2O.ai Driverless AI to test ML-driven inventory forecasting feasibility.
Python Azure Synapse Hive Impala H2O.ai

Healthcare Data Migration

Data integration and ETL improvements for a major healthcare client.

  • Managed data movement into Azure Data Lake Storage through updated Azure Data Factory pipelines.
  • Improved downstream availability by modifying Talend jobs and tuning SQL Server stored procedures.
Azure Data Factory Azure Data Lake Talend SQL Server

Selected GitHub work

Applied projects across AI systems and data engineering.

Hands-on work across retrieval, local model execution, guardrails, notebooks, containers, and API-backed applications.

RAG application Python

Multi-User RAG Chatbot

A containerized chatbot with FastAPI, Streamlit, Ollama, LangChain, LangGraph, PostgreSQL, Chroma, authentication, document processing, and tracing.

Open repository
AI safety Notebook

GuardRails

Exploration of model guardrails and safety checks, useful for building assistants that need predictable answers and clear boundaries.

Open repository
Deep learning Research

Deep Learning Repositories

A broad working collection across notebooks, Python, CUDA, C++, shell tooling, Docker, and learning material for modern model development.

Open repository
Agent framework Python

Microsoft AI Framework Demo

Sample project exploring Microsoft's agent framework and practical patterns for composing AI behavior in software.

Open repository
Document AI Python

Receipts Processor

A compact Python and Docker project for extracting useful information from receipt documents, showing practical automation around everyday data.

Open repository
Profile 135 repos

GitHub Portfolio

Public work spanning Python, notebooks, data structures, cloud learning, AI systems, and software engineering practice.

Browse all work

Technical range

Comfortable across model work, data systems, and product delivery.

Core tools used across model development, data pipelines, cloud delivery, and application engineering.

AI and Machine Learning

  • Machine Learning
  • Deep Learning
  • Generative AI
  • NLP
  • Computer Vision
  • Data Science

Model Tools

  • PyTorch
  • Scikit-learn
  • OpenCV
  • MLflow
  • LangChain
  • LlamaIndex

Data Engineering

  • SQL
  • Pandas
  • NumPy
  • Spark
  • Hadoop
  • Kafka
  • Redis

Cloud and Delivery

  • Docker
  • Kubernetes
  • ArgoCD
  • AWS
  • Azure
  • GCP
  • CI/CD

Application Engineering

  • Python
  • Go
  • REST APIs
  • Flask
  • React
  • Angular
  • JavaScript

Core Engineering

  • Bash
  • Git
  • Java
  • C
  • HTML
  • Jupyter

Education

Engineering foundation behind the applied work.

Formal training in electrical, computer, electronics, and communication engineering supports the practical ML and systems work shown above.

University of Southern California

MS in Electrical and Computer Engineering. August 2021 to May 2023.

Jawaharlal Nehru Technological University, Hyderabad

BTech and MTech in Electronics and Communications Engineering. August 2015 to November 2020.

Contact

Let's build AI systems that are useful, reliable, and ready for real teams.

Open to conversations about machine learning engineering, AI workflow systems, RAG applications, forecasting platforms, and cloud-native model delivery.

Email Krishna Download resume Connect on LinkedIn