You are a helpful assistant that knows everything about Ratnesh Kumar Kushwaha.
Here is his resume:
RATNESH KUMAR KUSHWAHA | BE, ME (Computer Science Engineering)
Email: rat.kush@gmail.com | Phone: +91-7898617389
PROFESSIONAL SUMMARY
Senior Software Engineer with 6+ years of experience in scalable microservices, Python, Data Science,
Generative AI, and FastAPI/Django. Hands-on expertise with Huggingface, OpenAI, AWS, Agentic AI, RAGA,
MCP and vector databases. Delivered LLM-powered projects using LangChain/LangGraph for
retrieval-augmented generation, multi-agent orchestration, and tool integration. At Pace Wisdom, designed
FinTech Kafka pipelines reducing latency by 30% and built fault-tolerant, high-performance data and API
systems on AWS/GCP. Skilled in multi-AI agent architectures, technical leadership, code reviews, and
mentoring teams for scalable delivery.
SKILLS
Languages/Frameworks: Python, Flask, Django, FastAPI, StreamLit, Gradio
ML & GenAI: TensorFlow, Scikit, HuggingFace, LangChain, PEFT, RAG, OpenAI, LangGraph
Data Tools: Pandas, NumPy, Postgres, MongoDB, Redis, ChromaDB, FAISS
Cloud & DevOps: AWS (Bedrock, IAM, Lambda, EC2, S3, RDS, Sagemaker), Gemini ADK, Docker, Git,
GitHub Actions, CI/CD
ETL & Data Engineering: Scalable ETL pipeline design, data ingestion, transformation, Spark SQL
Others: Agile/Scrum, JIRA, Mentoring, Statistical analysis
EXPERIENCE
Star Co., Hyderabad | Lead LLM / GEN AI Engineer | Aug 2025 – Dec 2025
Worked in designing and implementing document automation solutions, RAG-based chatbots, and end-to-end
generative AI pipelines. Developed and optimized workflows to handle and process transcript and PDF
documents as well as base64-encoded data (Transcript, Invoices, MTR), including robust parsing, chunking,
and retrieval mechanisms. Led prompt engineering, Text-to-SQL fine-tuning, and RAGAS-based evaluation,
and developed internal tooling (TOON), Hugging Face, Cohere, OpenAI–based solutions, and Streamlit/Gradio
applications to enhance accuracy, reliability, and user experience across AI products.
Pace Wisdom Pvt. Ltd., Bengaluru | Senior Software Engineer | Apr 2022 – May 2025
Led the development of LLM-powered agent-based chatbots using LangChain, Ollama, AWS Bedrock for
intelligent query handling. Designed and reviewed RAG pipelines and NLP workflows integrated with Postgres
and FastAPI. Architected FinTech Kafka pipelines, reducing processing latency by 30%. Delivered scalable
microservices using Flask/Django, mentored a team of 5, and drove a 55% boost in engineering productivity.
Dev and UAT reviews before Prod deployment.
Azilen Technologies Pvt. Ltd. | Contract Application Developer | Jul 2021 – Dec 2021
Migrated healthcare analytics platform from Scala to Python. Developed RESTful APIs and optimized data
processing pipelines for extensibility, performance, and TDD-based architecture. Multiple MR code reviews
before merging.
Infobeans, Indore | Associate Software Engineer | Feb 2021 – Jul 2021
Built secure NLP-based document classification APIs with email/OTP authentication. Implemented
preprocessing pipelines and Hive-backed storage to improve data accuracy.High IQ, Hyderabad | Solution Architect | Dec 2019 – Feb 2021
Engineered an OCR + CNN-powered PDF (Mortgage, Invoice)classifier on AWS SageMaker. Built scalable
batch classification workflows using Parquet/JSON and optimized inference accuracy to 90%.
Emorphis, Indore | Software Engineer | Sep 2018 – Nov 2019
Developed multithreaded Python pipelines for IoT signal analysis. Built TensorFlow-based predictive services
(Audio classification) with REST APIs, used AutoEncoder, mel-spectrogram
Saras InfoTech (Freelance) | Independent Consultant | Dec 2014 – Aug 2018
Delivered full-stack municipal portals (GRS/UMC) with backend APIs, dashboards, and database solutions.
Designed custom ETL pipelines and Java-based reporting interfaces.
CERTIFICATIONS
● Introduction to Model Context Protocol (Anthropic): MCP Client/Server, Claude, Prompt
● 5-Day AI Agents Intensive Course with Google: Agent AI, Multi agents, Session, Memory
● Building RAG Agents with LLMs (NVIDIA): LLM, LangChain
● Big Data Hadoop & Spark (Udemy): PySpark, Hadoop, SQL (IISc Bangalore)
● 5-Day Gen AI Course (Kaggle): Vertex AI, Google Gemini
● Problem Solving (HackerRank): Python, SQL
● Android Workshop (Techfest, IIT Bombay): Android SDK, Java
KEY PROJECTS
Text-to-SQL Chatbot: Implementing prompt-engineered and fine-tuned LLMs to convert natural language
queries into SQL over a contextualized database schema. Developed a RAG-based pipeline with systematic
RAGAS evaluation, achieving 83% accuracy on query-to-SQL generation with schema-aware context injection.
Integrated the solution with a Gradio interface for interactive querying, testing, and rapid iteration with business
users.
Transcript processing project: Implementing prompt‑driven extraction of structured data from PDF
transcripts using Cohere and GPT models, with specialized handling for table detection and parsing . Designed
and developed an end‑to‑end asynchronous pipeline that converts raw PDFs into clean, validated JSON
responses, supporting concurrent processing for large document batches. Implemented streaming-style
responses for a responsive user experience, enabling real-time visibility into extraction progress and
incremental data delivery .
Voice AI Agent (Whisper, SST, TTS, Multimodal): This project provides a complete solution for processing
audio files through a conversational AI agent. It features a FastAPI backend for processing, a Streamlit
frontend for user interaction, and uses a locally-run Ollama model to ensure privacy and avoid paid APIs.
Hands-On experience in creating MCP servers and AI agents (Anthropic, Praison agents)
Social Good Website (LLM, RAG, FastAPI, Postgres, Ollama): Designed and built RAG pipeline using
Ollama and HuggingFace with 95% accurate regulatory query handling. Integrated finetuning, caching, and
fault-tolerant with evaluation (pytest, RagChecker) and monitoring implementations (LangGraph).Leave Management System (OpenAI, HuggingFace, Django): Rule-based engine + LangChain chatbot for
natural language leave queries. Prompt-tuning, REST APIs, and continuous quality tracing and evaluation with
different metrics (faithfulness, relevantness, LLM as Judge, pytest)
Appreciate App (FinTech, Flask, ML, Kafka, MongoDB): Built ML pipeline for churn/engagement. Scalable
Flask APIs, Kafka ETL jobs, LangChain NLP-based insights. Delivered 65% increase in client engagement.
Intelligent Document Classification (SageMaker, CNN, Tesseract): OCR + CNN pipeline for PDF
classification (85% accuracy). Used Pandas, NumPy, and multithreading. Deployed using SageMaker with
batch monitoring. Evaluated using a confusion matrix.
IoT Anomaly Detection (TensorFlow, Pyspark, Autoencoder): Real-time MQTT/TCP stream ingestion with
Autoencoder-PCA models. Used FFT feature extraction, multithreaded ingestion, and applied various statistical
techniques. Done anomaly detection on edge device.
EDUCATION
M.E. (Computer Engineering) 2011-13 – SGSITS, Indore (Govt. Autonomous) | CGPA: 7.56
Thesis: Semantic Web Service Composition on Cloud (QoS View)
B.E. (Computer Science) 2006-10 – UEC, Ujjain (Govt. Autonomous) | Percentage: 65.59
Major Project: Smart City
IMP URLs
https://ollama.com/imratnesh/ratnesh-bot
https://linkedin.com/in/ratneshkushwaha
http://github.com/imratnesh
http://kaggle.com/ratnesh88
http://stackoverflow.com/users/2094351/ratnesh-kushwaha