About krishna
Most AI engineers can build a model. I build the system around it.I’m an AI/ML Engineer specializing in production-grade GenAI backends — RAG pipelines, LLM fine-tuning, and the FastAPI infrastructure that makes them run reliably at scale. Currently at Steves AI Lab, I architect systems that go from prototype to deployed API.What I've built: * High-accuracy RAG pipelines (LangChain + FAISS/ChromaDB) with semantic chunking that reduced hallucination rates in QA tasks * Fine-tuned Mistral 7B and LLaMA variants using QLoRA + PEFT on domain-specific datasets, then quantized with GGUF for CPU inference * Async FastAPI services containerized with Docker, handling concurrent LLM inference with sub-500ms P95 latency * MLOps pipelines with Apache Airflow + W&B experiment tracking, cutting model deployment cycle from days to hoursMy edge: I treat AI systems as engineering problems, not research experiments. Every pipeline I build is observable, reproducible, and designed to fail gracefully.Currently focused on:Agentic RAG systems (LangGraph) · LLM evaluation (RAGAS) · Distributed inference · Local LLMs with OllamaOpen to AI Engineer / Applied AI / Forward Deployed Engineer roles at early-stage AI-first startups.Let's build something that ships. — [email protected]
Key Skills
Experience
Ai / Ml Engineer
CurrentSteves Ai Lab
* Architected scalable AI APIs using FastAPI with async processing—reduced model inference latency by ~40% compared to synchronous Flask baselines on the same hardware * Engineered NLP pipelines with LangChain + OpenAI/GROQ APIs featuring semantic search (FAISS), improving retrieval relevance scores by ~28% across 3 client projects * Containerized 4 AI services with Docker, enabling one-command deployment and eliminating "works on my machine" failures across dev/staging/prod environments * Integrated RabbitMQ for async task queueing in a recommendation engine, decoupling the inference layer from the API layer and improving throughput under load
Data Science Intern
Techieshubhdeep It Solutions Pvt Ltd
* Fine-tuned Mistral 7B and LLaMA variants using QLoRA and PEFT on domain-specific datasets, optimizing hyperparameter configurations to minimize training loss * Deployed quantized LLM variants using GGUF and Ollama for localized, low-memory CPU inference pipelines, reducing cloud infrastructure dependency * Built and evaluated advanced Retrieval-Augmented Generation (RAG) architectures utilizing LangGraph and ChromaDB with RAGAS metrics for automated hallucination tracking * Streamlit UI development and modular Python packaging for cross-dataset generalization benchmarking and research-focused data visualization
Data Science Trainee
Almabetter
* Performed forensic data analysis on complex multi-table datasets, surfacing anomalies and producing audit-ready reports aligned with Deloitte consulting standards * Built end-to-end ML pipelines for classification and regression tasks, achieving 87%+ accuracy on 3 capstone projects across domains: finance, HR, and e-commerce * Delivered data-driven business insights for high-stakes scenarios, developing the ability to bridge technical findings and executive decision-making
Education
Shriram Group Of Colleges
Bachelors
Shriram Group Of Colleges
Bachelors
Interested in connecting with krishna?
Sign up for NinjaHire to send a connection request.
Common Questions
What is krishna's expertise?
krishna specializes in Fine-Tuning Engineer, with expertise in ai agent, amazon web services, analytical thinking, aws, big data.
Where is krishna located?
krishna is based in gwalior, madhya pradesh, india.
How much experience does krishna have?
krishna has 2+ years of professional experience.
How can I contact krishna?
You can connect with krishna through NinjaHire by signing up for a free account.
Other Fine-Tuning Engineers
Looking for a different Fine-Tuning Engineer?
Describe exactly who you need and NinjaHire will source them for you.
Type a role to try NinjaHire for free