Headline

Machine learning engineer

About

AI/ML Engineer with 2+ years of experience building production-grade AI systems. I specialize in Generative AI, Agentic AI, LLM applications, Retrieval-Augmented Generation (RAG), and intelligent automation. Currently at Finarkein Analytics, I build AI-powered document intelligence solutions for financial workflows, including OCR, information extraction, fraud detection, and high-throughput processing systems. Previously at Arya.ag, I developed AI agents, RAG-based chatbots, and Document AI solutions that automated business processes and reduced operational costs. My work has helped reduce OCR vendor dependency by 90%, automate planning workflows by 70%, and deflect 85% of routine support queries through AI-powered systems. I enjoy taking products from concept to production, designing scalable architectures, and solving real-world problems using AI. My technical stack includes Python, FastAPI, LangChain, LangGraph, CrewAI, Kafka, Docker, PostgreSQL, Vector Databases, and modern LLMs including OpenAI, Gemini, Llama, Mistral, Gemma, and Qwen. I'm passionate about building intelligent systems that can reason, retrieve knowledge, make decisions, and create measurable business impact.

Location

Pune District

AI stack

anthropic apis

antigravity

claude code

claude

codex

cursor

deepseek

gemini

geminicli

github copilot

hugging face

kimi

langgraph

langsmith

meta

n8n

notion

ollama

openai

openclaw

perplexity

Work experience

Machine Learning Engineer at Finarkein, Nov 2025 - Present · 9 mos, Developed a scalable multi-bank statement parser (bank statements, GST, ITR) supporting both text and complex image-based formats using custom regex pipelines, PaddleOCR, and open-source LLMs; built production-grade FastAPI microservice backed by Kafka for high-throughput scaling, webhooks for event-driven processing, and automated database logging of edge cases to enable continuous model accuracy improvement. Engineered SMS analytics system using regex-first feature extraction with Llama 3.2, tiny t5(seq2seq) fallback for missing critical fields.

Data Scientist at Arya.ag, Mar 2024 - Nov 2025 · 1 yr 9 mos, Noida, Uttar Pradesh, India, Built Document AI API using Gemini 2.0 Flash-Lite and FastAPI, processing documents at 1 paise per page. Created task reminder agent integrating Gmail, Zoho, and Bitbucket APIs with Airflow. Developed AI agents using OpenAI, LangChain, and Langgraph to help field teams plan and optimize their daily activities, reducing manual scheduling overhead. Built a conversational AI chatbot serving as the first point of contact for both business teams and customers, deflecting routine queries before escalation to human agents. Contributed to AryaQ grain quality application using CNN, YOLO, and DenseNet. actively involved in the Prakshep project, where I conduct data preprocessing and cleaning on satellite images of farms across multiple bands. Utilizing CNNs for analysis, I develop dashboards to estimate agricultural parameters such as stress, yield, and farm boundary analysis for diverse clients. build sustainability farming dashboard

Associate software developer Trainee at KPIT, Dec 2023 - Mar 2024 · 4 mos, Bengaluru, Karnataka, India, Gained insight to the software development, code debugging, scrum. Attended in-service training classes for C, AutoSar, Data Structure and algorithm, Agile, Automative industry.

Advanced robotics process automation at HighRadius, Jul 2023 - Nov 2023 · 5 mos, Bhubaneswar, Odisha, India, Working in advanced robotics process automation team

Summer Trainee at HighRadius, May 2023 - Jul 2023 · 3 mos, Remote, Built an AI Enabled Fintech B2B application to predict order amount that a customer might place in upcoming future, used various tools like data analytics,data cleaning, feature engineering,machine learning models, hyperparameter tuning, Java, api, javascript

Education

Bachelor of Technology, Computational Science, KIIT - Kalinga Institute of Industrial Technology

Public links

github: https://github.com/Saksham700