Available for full-time ML / NLP roles

Ankush Patil
ML & NLP Engineer

Building scalable LLM systems & production RAG pipelines · Agentic AI · Deep NLP

8+ end-to-end ML/NLP systems | 57K+ document RAG system | Deployed FastAPI inference APIs

PyTorch Transformers LLM Fine-tuning RAG LangChain LangGraph CrewAI n8n FastAPI Hugging Face FAISS Docker
Download Resume Get in Touch View Projects
I'm a Machine Learning and NLP engineer focused on building production-ready LLM systems, retrieval-augmented pipelines and efficient deep-learning models. Proficient in Python and C++; I have hands-on experience with GenAI and Agentic AI using LangChain, LangGraph, CrewAI and n8n — including multi-agent systems development and deployment. I deploy inference services via FastAPI and Docker. Currently targeting roles in NLP / LLM / AI engineering.

Experience

Jun 2024 – Present
ML / NLP Engineer – Self-Directed Projects
Self-driven Applied Research · Remote
  • Built end-to-end production systems: Legal RAG (EUR-Lex), LLaMA LoRA fine-tuning, custom Transformer models, multimodal CV classifiers.
  • Developed Agentic AI pipelines using LangChain and LangGraph; deployed FastAPI inference services with Docker.
  • Designed scalable workflows for data processing, training, evaluation (BLEU, ROUGE, Recall@K, MRR) and model hosting on Hugging Face Hub.
  • Strong expertise in PyTorch, Transformers, LLM fine-tuning, vector databases (FAISS), and classical ML pipelines.
Feb 2023 – Present
Site Engineer — Charwak Engineers & Developers
Onsite Engineering Role
  • Supervised onsite construction activities ensuring compliance with technical drawings, specifications, and safety standards.
  • Implemented documentation workflows and reporting processes, improving traceability and cross-team coordination.
  • Developed structured problem-solving, planning, and cross-functional coordination — skills directly transferable to ML system design.
Jul 2021 – Sep 2024
UPSC CSE & SSC CGL — Full-Time Preparation
Analytical & Reasoning-Focused Learning
  • Prepared for UPSC CSE 2022 & 2023, building analytical reasoning, structured thinking and deep comprehension skills.
  • Appeared for SSC CGL 2023 & 2024, sharpening quantitative aptitude, logical reasoning, and data-driven decision-making.
  • Developed disciplined self-learning (8–10 hrs/day) — strong planning, focus, and long-term goal management directly applicable to ML research.
2017 – 2021
B.E. Civil Engineering
K.K. Wagh Institute of Engineering Education & Research, Nashik

Projects

LLM
LLM
Wikipedia → LLaMA Fine-Tuning (RAG + LoRA)

Complete RAG pipeline with chunking, embeddings, FAISS retrieval and LoRA fine-tuning of LLaMA for factual QA. ROUGE-1: 0.033 • ROUGE-L: 0.024.

GitHub Sep–Oct 2025
NLP
NLP
Transformer — English ↔ Hindi (From Scratch)

Full Transformer encoder–decoder with custom attention, masking and beam search. BLEU 49.76.

GitHub Aug–Sep 2025
NLP
NLP
Amazon Reviews NLP Analysis (3M+ Reviews)

Scalable pipeline including sentiment, aspect-mining and BART/T5 summarization for millions of reviews.

GitHub Jul–Aug 2025
NLP
NLP
SMS Spam Classification — Naive Bayes, GDA, Logistic Regression

TF-IDF + classical models identifying high-impact spam tokens with strong recall on noisy text.

GitHub Feb–Mar 2025
CV
Computer Vision
Plant Disease Detection — ResNet18

ResNet18 transfer-learning pipeline with Grad-CAM explainability; high accuracy across 38 classes.

GitHub Apr–Jun 2025
ML
Classical ML
Heart Stroke Prediction System

End-to-end ML pipeline for stroke risk prediction with feature engineering and class imbalance handling.

GitHub 2025
ML
Classical ML
Credit Card Fraud Detection — Anomaly Models

Hybrid anomaly-detection workflow (PCA, UMAP, Isolation Forest). Isolation Forest delivered best precision-recall.

GitHub Feb–Mar 2025
ML
Classical ML
Student Performance & Ranking System

Stacked regression with PCA, SVM, and Gradient Boosting. R² = 0.90.

GitHub Jan–Feb 2025

Skills & Expertise

📊
Machine Learning Algorithms
Linear Regression Logistic Regression SVM Decision Trees Random Forest Gradient Boosting XGBoost K-Means PCA Isolation Forest Naive Bayes GDA
🧠
Deep Learning
PyTorch CNN RNN LSTM GRU Transformers Attention Mechanism Encoder–Decoder Transfer Learning Grad-CAM
📝
Natural Language Processing
TF-IDF BoW / n-grams Tokenization Seq2Seq Models Beam Search Topic Modeling Sentiment Analysis BLEU / ROUGE BART / T5 Text Summarization
👁️
Computer Vision
ResNet Architectures Image Classification Transfer Learning (CV) Grad-CAM Explainability Data Augmentation OpenCV Basics
🤖
LLMs & Generative AI
LLaMA Models Hugging Face Transformers SentenceTransformers LLM Fine-tuning (LoRA / PEFT) Prompt Engineering Multimodal LLMs Agentic AI Workflows
🔍
RAG & Retrieval Systems
FAISS Vector Search Dense Retrieval Sparse Retrieval Hybrid RAG Pipelines Chunking Strategies Reranking Recall@K / Precision@K MRR Evaluation
⚡
Agentic AI Frameworks
LangChain LangGraph CrewAI n8n Multi-Agent Systems Development
⚙️
Tools, Languages & Infrastructure
Python C++ FastAPI Docker REST APIs Git / GitHub Structured Logging Hugging Face Hub NumPy / Pandas Scikit-learn

Now Learning

Staying current with the frontier of AI engineering.

Agentic AI Systems — multi-step agents that plan, reason & execute
Multimodal LLMs — image + text fusion architectures
Active Project — Automating real-world workflows using Agentic AI

Let's Connect

Open to full-time ML / NLP / AI roles

I'm actively looking for opportunities where I can build scalable, real-world AI systems. Reach out for interviews, collaborations, or just to talk ML.

📧 Email Me 🔗 LinkedIn Profile 🐙 GitHub — Ankush-Patil99
Contact Me for Interviews or Collaborations →
↑