Arnab Mitra — Software Engineer

🚀 Building AI & ML Systems | LLMs, Data Pipelines, Cloud (AWS, Azure), ETL, MLOps I build AI solutions that solve real-world problems and optimize data systems to make them scalable and efficient. With experience in machine learning, cloud-based ETL pipelines, and AI automation, I specialize in bridging the gap between raw data and actionable insights. Currently pursuing my M.S. in Data Science at Arizona State University, I am actively seeking Summer 2025 Data Science internships to apply my expertise in AI and data engineering. Key Achievements & Impact 🔹 Cloud & Data Engineering: Designed and optimized a star schema database for Indigo Airlines, improving query performance by 20% and streamlining cloud migration. 🤖 AI-Powered Automation: Developed and deployed a GPT-3.5-powered HR chatbot on Azure, implementing a Retrieval-Augmented Generation (RAG) pipeline that reduced HR workload by 40% while ensuring automated monitoring and alerts. 📈 Forecasting & Data Pipelines: Built an AWS-enabled CI/CD forecasting pipeline, automating data ingestion, preprocessing, and model training for high-accuracy predictions. ⚡ Data Analytics & Optimization: Enhanced Nestlé’s Power BI reports, reducing refresh times by 50% through advanced SQL query optimization across 50+ tables, accelerating data-driven decisions. 🌍 Accessible AI for Social Good: Developed an AI-powered image-to-speech conversion system for visually impaired users, leveraging ResNet50 for image feature extraction and LSTM with GloVe embeddings for contextual text generation. Technical Expertise ✅ Machine Learning & AI: TensorFlow, PyTorch, Scikit-learn, NLP, LLMs, multimodal architectures ✅ Cloud & DevOps: AWS (Lambda, S3, Glue), Azure (Data Factory, SQL, Application Insights), CI/CD, Docker, MLFlow ✅ Data Engineering & ETL: Scalable ETL pipelines, data lakes & warehouses, data versioning with DVC ✅ Programming & Tools: Python, SQL, Unix/Linux, Flask API deployment ✅ Data Visualization & Analytics: Tableau, SQL-based reporting, Excel Publication 📄 "A Comparative Study of Demand Forecasting Models for a Multi-Channel Retail Company" (Springer) – A novel hybrid ML approach for accurate retail forecasting. 🚀 I am passionate about leveraging AI, automation, and cloud computing to build scalable, high-impact data solutions. Let’s connect! 📧 arnabmitra410@gmail.com | 📞 +1 (623) 286-5758

Stackforce AI infers this person is a Data Engineer and AI Specialist with a focus on SaaS and cloud solutions.

Location: Tempe, Arizona, United States

Experience: 2 yrs 2 mos

Skills

Ai & Ml
Cloud Computing
Data Engineering
Data Analytics
Accessibility Technology

Career Highlights

Designed a star schema database improving query performance by 20%.
Developed a GPT-3.5-powered HR chatbot reducing workload by 40%.
Built an AWS-enabled CI/CD forecasting pipeline for high accuracy.

Work Experience

Gen

Software Engineer (9 mos)

Software Engineer Intern (3 mos)

Arizona State University

Operations Support Lead (1 mo)

Operations Support Specialist (6 mos)

TransOrg Analytics

Data Analyst (1 yr 8 mos)

Teksands.ai

Data Science Intern (4 mos)

Education

Master's degree at Arizona State University

Bachelor of Technology - BTech at Delhi Technological University (Formerly DCE)

Arnab Mitra

Software Engineer

Tempe, Arizona, United States2 yrs 2 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Designed a star schema database improving query performance by 20%.
Developed a GPT-3.5-powered HR chatbot reducing workload by 40%.
Built an AWS-enabled CI/CD forecasting pipeline for high accuracy.

Stackforce AI infers this person is a Data Engineer and AI Specialist with a focus on SaaS and cloud solutions.

Contact

arnabmitra410@gmail.com LinkedIn

Skills

Core Skills

Ai & MlCloud ComputingData EngineeringData AnalyticsAccessibility Technology

Other Skills

AI Classification SystemsAI SolutionsAI Taxonomy ExpansionAWSAWS CloudFormationAWS GlueAWS LambdaAWS SageMakerAccessibility SolutionsAlgorithm DesignAlgorithmsAmazon S3Amazon Web Services (AWS)AnacondaApplied Machine Learning

About

Experience

2 yrs 2 mos

Total Experience

1 yr

Average Tenure

Current Experience

Gen

2 roles

Software Engineer

Sep 2025 – Present · 9 mos · Tempe, Arizona, United States · Hybrid

Software Engineer Intern

Jun 2025 – Sep 2025 · 3 mos · Tempe, Arizona, United States · Hybrid

🚀 Built and deployed AI-powered classification systems using large language models, prompt engineering, and cloud-native microservices to automate Salesforce case handling and categorize NPS survey comments.
Key Achievements:
1. Salesforce Case Classification
Engineered a Ruby-based microservice integrated with Google Gemini Flash Lite 2.0 to automate classification of Salesforce support cases, with an accuracy of 88% by combining AI outputs with agent notes.
Designed a two-step classification pipeline with 16 specialized prompts, reducing hallucinations and ensuring schema-controlled outputs for Issue Types, Subtypes, and Categories.
Optimized performance to deliver predictions in <950ms per case at $0.00035 per case, enabling scalability across 39K+ monthly chat cases for <$15/month.
Built and deployed with Docker + Kubernetes (GKE), secured using Google Cloud IAM, and integrated into Salesforce workflows (via Romulus) using REST APIs with validation and robust error handling.
Conducted a BigQuery-driven POC using 1,400+ historical cases to benchmark Gemini performance before production rollout.
2. NPS Survey Comment Categorization
Developed a BigQuery + Gemini pipeline to categorize Qualtrics NPS survey comments into 10 predefined topics (e.g., Pricing/Value, Customer Service, Messaging).
Authored prompt-engineered SQL queries with strict schema definitions to ensure parseable and consistent classification results.
Designed an AI-driven taxonomy expansion workflow that surfaced 339 new recurring themes (e.g., Website Usability, Language Barriers), extending coverage of customer feedback.
Delivered a sandbox environment in BigQuery for stakeholders to validate category assignments and explore taxonomy recommendations.
Technical Stack:
Backend & APIs: Ruby, Sinatra, Puma, REST APIs, JSON Web Token
Cloud & Deployment: Docker, Kubernetes, Google Cloud IAM, GitHub Actions, Teamcity, Terraform, Big Query, SQL

Large Language ModelsCloud-native MicroservicesPrompt EngineeringSalesforce IntegrationAI Classification SystemsAI & ML+1

Arizona state university

2 roles

Operations Support Lead

May 2025 – Jun 2025 · 1 mo · Tempe, Arizona, United States · On-site

Operations Support Specialist

Nov 2024 – May 2025 · 6 mos · Tempe, Arizona, United States · On-site

🚀 Supporting Admission Services at ASU by managing graduate and transfer student enrollment deposits and application fees, ensuring a seamless and efficient enrollment process.
Key Responsibilities & Achievements:
Case Management & Student Support: Handle high-volume student queries related to enrollment deposits and application fees via Salesforce, ensuring timely resolution and exceptional applicant experience.
Process Optimization: Streamlined case tracking and resolution workflows, improving operational efficiency by 15% and enhancing workflow transparency.
Data Management & Reporting: Developed automated data cleaning workflows in Excel (advanced formulas, pivot tables, macros) to standardize datasets, identify trends, and improve reporting accuracy.
Front Desk Operations: Occasionally manage the Admission Services front desk, assisting students in person and providing critical guidance on enrollment-related processes.
Technical Stack:
CRM & ERP Platforms: Salesforce, PeopleSoft
Data Tools: Microsoft Excel (advanced formulas, pivot tables, macros)
Analytics & Reporting: Data validation, cleansing, and preparation for analysis
🏆 Recognized for proactive problem-solving and data accuracy, contributing to a seamless enrollment experience for students.

Transorg analytics

Data Analyst

Jun 2022 – Feb 2024 · 1 yr 8 mos · Gurgaon · Hybrid

🚀 Leveraged data engineering and analytics to optimize cloud databases, automate forecasting pipelines, and enhance business intelligence for aviation, FMCG, and internal projects.
Key Achievements:
1. Cloud Optimization for Indigo Airlines: Designed a star schema for 100+ tables during cloud migration, improving query performance and creating a data dictionary for seamless stakeholder communication. Presented the optimized model to leadership, driving strategic decisions.
2. AI-Powered HR Chatbot Deployment: Built and deployed a GPT-3.5-based HR chatbot on Microsoft Azure using a Retrieval-Augmented Generation (RAG) pipeline. Integrated logging and automated misuse alerts, enhancing HR query efficiency.
3. Power BI Performance Enhancement for Nestlé: Reduced report refresh time by 50% by optimizing SQL queries across 50+ tables, accelerating daily reporting and decision-making.
4. Azure ETL Pipeline Redesign: Rebuilt Azure Data Factory pipelines for Nestlé, eliminating inefficiencies, automating data ingestion, and reducing run times for improved data accuracy.
5. Automated Forecasting Pipeline: Developed an end-to-end, CI/CD-enabled forecasting pipeline on AWS, automating data preprocessing, model training, and deployment for increased accuracy and efficiency.
Technical Stack:
1. Cloud & DevOps: Azure (Data Factory, SQL, Application Insights), AWS (Lambda, S3, RDS)
2. Data Engineering & Analytics: SQL (Optimization), Power BI, Python (Pandas, NumPy)
3. AI & NLP: GPT-3.5 Turbo, RAG Pipelines, OpenAI API
4. CI/CD: Jenkins, GitHub Actions
🏆 Recognized by leadership for driving seamless cloud migrations and delivering AI-powered HR solutions.

Cloud OptimizationData EngineeringAI SolutionsSQL OptimizationETL PipelinesAI & ML

Teksands.ai

Data Science Intern

Oct 2020 – Feb 2021 · 4 mos · Remote

🚀 Developed AI-driven accessibility solutions, leveraging deep learning and NLP to enhance visual content accessibility for visually impaired users.
Key Achievements:
1. Engineered an AI-powered descriptive audio system that converts images into spoken narratives, improving real-time access to visual content.
2. Built a multimodal AI framework integrating ResNet50 for image feature extraction and LSTM for text generation, optimized with GloVe embeddings, leading to fluent and contextually accurate captions.
3. Designed a custom data generator to streamline model training, reducing processing time and improving neural network efficiency.
4. Developed a robust text preprocessing pipeline with tokenization, stemming, stopword removal, and <start>/<end> tokens, enhancing caption generation quality and training accuracy.
5. Integrated Google Text-to-Speech (gTTS) for real-time audio conversion, enabling seamless accessibility for diverse users.
Technical Stack:
1. Machine Learning & Deep Learning: TensorFlow, Keras, ResNet50, LSTM, GloVe
2. Programming & Data Processing: Python, Pandas, NumPy
3. Text Processing: Tokenization, stemming, stopword removal
4. Audio Processing: gTTS for real-time text-to-speech conversion
4. Optimization: Custom data generators, model tuning
🏆 Contributed to the development of an AI-driven accessibility platform, enhancing inclusivity for visually impaired users through real-time multimodal interactions.