AI/ML Engineer focused on building practical Machine Learning and LLM-powered applications using Python, FastAPI, SQL, and NLP workflows.

• Generative AI & LLM Applications
• NLP & Prompt Engineering
• Machine Learning Pipelines
• FastAPI Backend Development
• SQL & Database Workflows
• Real-world AI Problem Solving

Languages & Databases
Python • SQL • SQLite

AI/ML
Scikit-learn • XGBoost • TensorFlow • Machine Learning Pipelines

Generative AI & NLP
LLMs • Prompt Engineering • LangChain • RAG • FAISS • Groq API

Backend & Deployment
FastAPI • Flask • REST APIs • Gradio • Hugging Face Spaces

Data & Analytics
Pandas • NumPy • Feature Engineering • Data Preprocessing

Developed an NL-to-SQL application using FastAPI, Groq API, and SQLite to convert natural language questions into SQL queries and structured database responses.

Key Highlights

• Implemented query validation logic for safer SQL execution
• Handled edge cases related to SQLite syntax compatibility
• Improved query generation consistency using prompt refinement
• Built modular backend APIs for interactive analytics workflows
• Focused on debugging and reliability of LLM-generated SQL outputs

Contributed a merged example: a local RAG pipeline (LangChain, FAISS, FLAN-T5) validated with BOOTH’s check_with_evidence() and a custom NLI-based comparison function (DeBERTa cross-encoder). Documented the failure modes honestly, including false blocks on terse answers, and addressed maintainer review feedback.

Technologies: Python, LangChain, FAISS, FLAN-T5, DeBERTa, NLI, RAG

Pull Request: BOOTH PR #22 — Merged Contribution

• LinkedIn: linkedin.com/in/harithalakshmikt097
• GitHub: github.com/Harithalakshmikt
• Email: harithalakhmi097@gmail.com