Open to opportunities

Krishna Nivja

@krishnanivja

Message

AI/ML engineer building real-time agentic RAG and fine-tuned LLM systems.

India

Message

What I'm looking for

I’m looking for an AI/ML role where I can build production RAG/LLM systems, reduce hallucinations, and ship fast APIs and deployed pipelines on AWS—where engineering quality, monitoring, and model efficiency are valued.

I’m an AI/ML Engineer focused on turning LLM capabilities into reliable, production-grade systems—especially with agentic RAG for real-time, context-aware responses. I enjoy bridging retrieval, reasoning, and deployment so models behave consistently in real user scenarios.

At Steve’s AI Lab, I built a GeoSpatial AI platform integrating Google Maps API, OpenStreetMap, and 5+ REST APIs with a LangChain RAG pipeline and ChromaDB for location-aware query resolution. I engineered a LangGraph-based agentic RAG system that reduced LLM hallucinations by 40% on geo-specific queries.

I also containerised and deployed the inference stack with Docker on AWS EC2, monitoring it with Weights & Biases and maintaining 99.2% production uptime. On the model side, I fine-tuned LLMs using LoRA/QLoRA (Hugging Face PEFT) with GGUF/GPTQ quantization to keep performance efficient.

Earlier, as an AI/ML Intern, I built ML pipelines with Python, Scikit-learn, and XGBoost, using Optuna to reduce overfitting. I designed CNN and BiLSTM models in PyTorch and TensorFlow, and I optimized NLP inference by 40% using model quantization and tokeniser caching, deploying a real-time FastAPI service.

Experience

Work history, roles, and key accomplishments

Current

AI/ML Engineer

Current

Steve’s AI Lab

Apr 2026 - Present (2 months)

Built a geospatial AI platform integrating Google Maps API, OpenStreetMap, and 5+ REST APIs with a LangChain RAG pipeline and ChromaDB for real-time, location-aware query resolution. Engineered agentic RAG with LangGraph, reducing LLM hallucinations by 40%, and deployed a Dockerized inference stack on AWS EC2 with 99.2% production uptime.

LangGraph ChromaDB Vector Databases Docker AWS EC2 Weights Biases

AI/ML Intern

Techieshubhdeep Pvt. Ltd.

Apr 2025 - Jan 2026 (9 months)

Developed ML pipelines in Python using Scikit-learn and XGBoost, applying feature engineering, cross-validation, and Optuna hyperparameter tuning to improve generalization. Designed CNN and BiLSTM models (PyTorch/TensorFlow) and optimized NLP inference by 40% using quantization and tokenizer caching, deploying a real-time FastAPI service to reduce per-request compute cost.

Python scikit learn XGBoost Optuna CNN PyTorch TensorFlow Model Quantization fastAPI Bi LSTM

Education

Degrees, certifications, and relevant coursework

Shri Ram Group of College

Bachelor of Technology, Artificial Intelligence and Machine Learning

2024 - 2028

Activities and societies: Coursework: ML, Deep Learning, NLP, Computer Vision, DSA, OS, DBMS, OOP. Certifications: ChatGPT Prompt Engineering for Developers (DeepLearning.AI), AWS Machine Learning Foundations (Udacity).

Pursuing a B.Tech in Artificial Intelligence and Machine Learning with coursework in machine learning, deep learning, NLP, and computer vision. Completed certificates in ChatGPT Prompt Engineering (DeepLearning.AI) and AWS Machine Learning Foundations (Udacity).