Looking for a job

koushik SALAMMAGARI

@koushiksalammagari

Message

Software Engineer at BalanceAI building multi-agent AI systems that process 10K+ sensor data points per query.

United States

Message

At BalanceAI, I design and ship generative AI systems with multi-agent workflows, processing 10K+ sensor data points per query. I also built agentic solutions for statistical analysis and deployed a Next.js frontend to AWS Amplify.

Previously, I shipped AI-powered call-routing features at AT&T that reduced resolution times by 45%, and developed RAG evaluation systems and fine-tuned language models. At Symantrix, I built full-stack machine-learning applications for document processing; my open-source contributions include production features and tests for Google Agent Development Kit and Giskard.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer (AI/ML)

Current

BalanceAI

Sep 2025 - Present (1 year)

Collaborated with product and engineering teams to design and ship generative AI systems with multi-agent workflows using LangGraph, processing 10K+ sensor data points per query with real-time response latency. Leveraged AI-assisted development tools to accelerate feature development by 40% and built agentic AI solutions for statistical computations, delivering insights identifying 15-35% optimiza

LangGraph Lang Chain GitHub Co Pilot Prompt Engineering Evaluation Frameworks Next.js AWS Amplify Real Time Data Streaming Conversational Memory

Software Engineer (GenAI)

AT&T

Sep 2024 - Jul 2025 (10 months)

Shipped AI-powered call routing features to production, collaborating with product managers and designers to translate requirements into generative AI systems with multi-agent workflows and LLM orchestration patterns using LangGraph and LangChain, reducing resolution times by 45%. Developed task-specific evaluation metrics and datasets for RAG systems, integrating vector databases for information

LangGraph Lang Chain Model Context Protocol RAG Faiss Pinecone QLoRA LoRAs PyTorch GPT 4 LLaMA 2 fastAPI A B Testing

Software Engineer (ML and Fullstack)

Symantrix

Aug 2020 - Aug 2022 (2 years)

Built full-stack ML applications for document processing with React frontend and Python backend, developing production NLP features for classification and semantic similarity using PyTorch and TensorFlow. Transitioned rule-based logic into classical ML pipelines for automated email triage and document tagging, and built supervised and unsupervised ML models to detect anomalies and forecast sales.

React Python PyTorch TensorFlow NLP Classification Semantic Similarity AWS S3 EC2 Sagemaker Random Forest K Means XGBoost

Co-Founder, FullStack Developer

Learnasky

May 2019 - Aug 2020 (1 year 3 months)

Developed full-stack web application using MERN stack, architected MongoDB schemas and built RESTful APIs with Node.js and React, improving workflow efficiency by 30% and reducing latency by 15%.

MERN Stack MongoDB NodeJS React REST APIs