Skip to content
View Athilesten's full-sized avatar
🎯
Focusing
🎯
Focusing
  • 20:10 (UTC -12:00)

Block or report Athilesten

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Athilesten/README.md

Harshit Saxena

Python Developer · Backend Engineering · Data Engineering · Applied AI/ML
Bengaluru, India

I build Python APIs, data pipelines, and AI applications. My projects cover transaction risk scoring, SQL analytics, and computer vision. I’m interested in engineering roles in Japan and in international teams working on Python backends, data platforms, or applied AI.

LinkedIn · Portfolio repository

Featured projects

Fraud Risk Intelligence Platform

Built a portfolio application that scores transactions with machine learning and rule-based checks. FastAPI serves the scores, PostgreSQL records decisions for audit, and Kafka and PySpark stream results into Delta Lake. Docker Compose runs the demo locally, with Prometheus and Grafana for monitoring.

Agentic SQL Data Analyst

Contributed to the LangChain agent, SQL validation, prompt design, evaluation, and demos for a C-DAC team project. The application uses Gemini to translate plain-English questions into SQL, validates queries before read-only execution in PostgreSQL, and can retry failed queries through a correction loop.

Journal-Entry Audit Analytics

Analyzed 7,500 synthetic journal entries using PostgreSQL audit rules, Benford’s Law, and Isolation Forest. The pipeline ranks exceptions for review and exports the results for Tableau.

Real-Time Object Detection

Built a Python application that runs a pretrained YOLO model on webcam feeds and video files. OpenCV overlays bounding boxes, confidence scores, object counts, and FPS. I tested how lighting, input size, and confidence thresholds affected detection and speed.

Technical skills

  • Languages and backend · Python, SQL, FastAPI, REST APIs, API integration
  • Data engineering and databases · Kafka, PySpark, ETL, Delta Lake, PostgreSQL, MySQL
  • Machine learning and generative AI · scikit-learn, PyTorch, TensorFlow, LangChain, LLMs, RAG, OpenCV, YOLO
  • Development and monitoring · Docker, Git, GitHub, Prometheus, Grafana

Education

  • PG Diploma in Big Data Analytics, C-DAC Bengaluru, 2026
  • Master of Computer Applications (MCA), Invertis University, Bareilly, 2025

Certifications and training

  • AWS · AI/ML Fundamentals, Generative AI on AWS
  • IBM · Python Fundamentals, Certified Enterprise Java Developer
  • Job simulations · Deloitte Data Analytics & Cyber, Tata Group Data Visualisation

Pinned Loading

  1. fraud-risk-intelligence-platform fraud-risk-intelligence-platform Public

    Fraud risk scoring with FastAPI, PostgreSQL audit logs, Kafka and PySpark streaming, and Prometheus/Grafana monitoring. Runs locally with Docker Compose.

    Python

  2. Agentic-SQL-Analyst-project Agentic-SQL-Analyst-project Public

    Natural-language SQL analytics with Python, LangChain, Gemini, and PostgreSQL. Validates queries for read-only execution and retries failures through a correction loop.

    Python

  3. journal-entry-audit-analytics journal-entry-audit-analytics Public

    Explainable journal-entry audit analytics using PostgreSQL, Python, Isolation Forest, Benford's Law, and Tableau-ready reporting.

    Python

  4. yolo-opencv-object-detector yolo-opencv-object-detector Public

    Real-time object detection with YOLO and OpenCV, including FPS, object counts, and video export.

    Python