COLLECTION '26 • AI SYSTEMS

Sahil Singh

A AI Engineer building in India.

AI Engineer with experience building production-grade Computer Vision, Generative AI, and Retrieval-Augmented Generation (RAG) systems. Skilled in developing LLM-powered applications, real-time video analytics pipelines, multimodal AI workflows, and scalable backend services using Python, PyTorch, FastAPI, AWS, and modern AI frameworks. Experienced in fine-tuning large language models, optimizing inference pipelines, and deploying AI-driven applications for production environments.

SCROLL

Engineering intelligence for high-performance edge computing.

I design computer vision models, spatiotemporal CCTV analysis grids, fine-tuned large language models, and enterprise knowledge indexing graphs (RAG). By focusing on serverless deployments and in-browser sandboxed processing, I build applications that are secure, private, and blisteringly fast.

sahil_singh.sys

// Diagnostic Report

Location: Gurugram, India (IST)

Active Node: Radiansys Inc.

Primary Core: AI & Computer Vision Pipelines

Active Pipelines: Running [100% OK]

> sahil.get_contact_info()

sahilsingh82738235@gmail.com

+91-7505085844

01.

Flagship Product

ACTIVE SAAS

InstaFuture.in — Unified Ecosystem

A production-grade utility suite engineered to handle high-volume file compilations, text rephrasing, and local development tools directly inside serverless edge runtimes.

Premium PDF workflows
Multi-language sandbox
Live AI writing systems
Launch Live Site ↗
📄
PDF WORKSPACE
PDF WORKSPACE

PDF Tools Page

PDF WORKSPACE InstaFuture.in Suite

PDF Tools Page

High-trust PDF flows for merge, sign, watermark, split, and conversion with a real desktop workspace feel.

✍️
AI WRITING LAB
AI STUDIO

AI Tools Page

AI STUDIO InstaFuture.in Suite

AI Tools Page

AI workflows for summarizing, rewriting, and polishing text with direct, task-focused premium interactions.

💻
UTILITIES / CODE
COMPILER LAB

Utilities Page

COMPILER LAB InstaFuture.in Suite

Utilities Page

Sandboxed compiler executions for Python and web previews, built to feel fast, visual, and approachable.

🔀
DATA FORGE
DATA FORGE

Data Tools Page

DATA FORGE InstaFuture.in Suite

Data Tools Page

Structured data workflows for JSON and CSV schemas wrapped in responsive interaction designs.

02.

Work Experience

Feb 2026 – Present Gurugram, India | Full-Time
  • Developed real-time computer vision pipelines for face detection, object detection, person tracking, and recognition using OpenCV and YOLOv26-based models.
  • Built AI-powered sports highlight generation systems using event detection, video analytics, and multimodal video understanding techniques.
  • Fine-tuned and evaluated multimodal and large language models (LLMs) including Qwen for domain-specific AI applications.
  • Implemented scalable AI inference workflows, REST APIs, and backend microservices using FastAPI, Docker, and AWS (Lambda, S3, SNS).
  • Worked on deep learning optimization and deployment pipelines for production AI systems.
PythonPyTorchTensorFlowOpenCVYOLOv26Object DetectionVideo AnalyticsTrackingFastAPIDockerQwenNLPDeep LearningGenerative AILLMsAWS

Machine Learning Engineer

Nexority
Aug 2025 – Feb 2026 Hybrid | Full-Time
  • Developed machine learning models for predictive analytics and classification tasks.
  • Performed feature engineering, exploratory data analysis, preprocessing, and model evaluation using Pandas, NumPy, and SQL databases.
  • Contributed to deployment and optimization workflows for ML-based business solutions.
  • Collaborated with teams to improve data-driven decision-making systems.
PythonSQLPandasNumPyScikit-learnMachine LearningCatBoostXGBoostLightGBM

Data Science Intern

KultureHire
Feb 2025 – May 2025 India (Remote) | Internship
  • Worked on real-world data analysis and machine learning projects using Python, SQL, Excel, and Power BI.
  • Performed data cleaning, preprocessing, exploratory data analysis (EDA), and statistical analysis to derive actionable insights.
  • Built interactive dashboards and visual reports in Power BI to support data-driven decision-making.
  • Applied feature engineering and statistical modeling techniques on public domain datasets for predictive analytics tasks.
  • Gained hands-on experience in predictive workflows using Pandas, NumPy, MySQL, and modern visualization libraries.
PythonSQLMySQLPandasNumPyExcelPower BIExploratory Data Analysis (EDA)Feature EngineeringStatistical Analysis
03.

Selected Works

A showcase of machine learning pipelines, RAG engines, video trackers, and neural network applications. Open any card to launch a larger workspace with live simulations, controls, and project-specific interactions.

Q4 02:14 AI DETECTED: DUNK + CROWD SURGE
TRANSCRIBE
EVENT SCORE
FFMPEG CUT
Project #1

AI Sports Highlight Generation

Project #1 Python, FFmpeg, OpenAI, AWS Lambda, S3, NestJS, MongoDB, Object Detection, Video Analytics

AI Sports Highlight Generation

An automated AI pipeline that processes full sports footage, detects highlights using multi-modal analysis, and generates clips.

Launch interactive workspace
ASD 0.982 FACE LOCK
TALKNET
FACE TRACK
STABILIZE
Project #2

Speaker-Aware Portrait Reframer

Project #2 Python, OpenCV, MediaPipe, PyTorch, TalkNet-ASD, YOLOv26, Object Detection, Video Analytics, Tracking

Speaker-Aware Portrait Reframer

An AI video cropping pipeline converting landscape footage into mobile aspect ratios by tracking active speakers.

Launch interactive workspace
SOP 4.2 FAQ 03 POLICY SPEC
refund policy after 30 days?
Context -> SOP_v2.pdf / Section 4.2 / score 0.941
Project #3

Enterprise Knowledge Assistant (RAG)

Project #3 FastAPI, LangChain, LlamaIndex, OpenAI, FAISS, PostgreSQL, Docker, RAG Pipelines, Embeddings, Vector Search, LLMs

Enterprise Knowledge Assistant (RAG)

A secure Retrieval-Augmented Generation assistant indexing SOPs and docs with source-grounded conversational citations.

Launch interactive workspace
show active users from north region with purchases above 500
SELECT name, region, amount FROM users JOIN purchases ON users.id = purchases.user_id WHERE users.active = true AND region = 'North' AND amount > 500;
Name
Region
Amount
Rahul
North
$890
Project #4

Domain-Specific SQL Assistant

Project #4 Python, OpenAI Fine-Tuning, PostgreSQL, FastAPI, LangChain, LLMs, REST APIs, SQL

Domain-Specific SQL Assistant

A fine-tuned LLM translating natural language queries to database queries with robust schema validations.

Launch interactive workspace
Generative AI & LLM Fine-Tuning 98.7%
Deep Learning Foundations & CNNs 92.1%
Data Science Tools & Operations 86.4%
Project #5

Smart AI Search Tool for Free Courses

Project #5 Python, NLP, Gradio, Vector Search, Hugging Face, Embeddings

Smart AI Search Tool for Free Courses

An NLP-powered semantic search engine utilizing custom embeddings and vector distance calculations to rank and query Analytics Vidhya datasets.

Launch interactive workspace
HQ D1 D2 D3 D4
3 vehicles / 102.2 km / OR-Tools
Project #6

TransLogi Delivery Route Optimization

Project #6 Python, Streamlit, OR-Tools, Predictive Analytics, Route Optimization, Pandas, NumPy

TransLogi Delivery Route Optimization

A comprehensive logistics system integrating predictive analytics, multi-vehicle route optimization, and real-time monitoring dashboards.

Launch interactive workspace
04.

Publications & Anomaly Research

Comparative Analysis of Suspicious Activity Detection Techniques in Surveillance Videos
IEEE ICDT 2025 Jan 2025 • 8 min read

Comparative Analysis of Suspicious Activity Detection Techniques in Surveillance Videos

  • In this study, we explore and evaluate various cutting-edge approaches to automated anomaly detection in surveillance videos — including CNN-based models, spatiotemporal autoencoders, and transfer learning techniques.
  • With public safety becoming increasingly reliant on real-time monitoring, we identify methods offering the optimal balance of accuracy, computational efficiency, and practical real-world deployment.
  • DOI: 10.1109/ICDT63985.2025.1098638
05.

Metrics & Achievements

1+ Years

Active Pipeline Engineering

10+ Pipelines

Architected & Deployed

0→1 Builder

Product + AI Delivery

2nd Rank in Data Science Hackathon

Secured 2nd rank among 100+ competitors at Sharda University, solving predictive modeling and feature extraction pipelines.

Published IEEE Research Paper

Authored and published computer vision research on anomaly detection in CCTV networks at IEEE ICDT 2025 conference.

Merit-Based Scholarship Recipient

Awarded prestigious academic scholarship twice by Sharda University for top academic standing and GPA of 8.08/10.

06.

Technical Expertise

Programming

Python
SQL

Generative AI

LLMs
RAG Pipelines
Prompt Engineering
Embeddings
Vector Search (FAISS/Chroma)
LangChain
LlamaIndex
OpenAI APIs

Computer Vision

OpenCV
YOLOv26
MediaPipe
Object Detection
Video Analytics
Tracking
Active Speaker Detection (TalkNet-ASD)

AI / ML Frameworks

PyTorch
TensorFlow
Scikit-learn
CatBoost
XGBoost
LightGBM
Deep Learning

Backend & Cloud

FastAPI
Docker
AWS (Lambda, S3, SNS)
PostgreSQL
MongoDB
NestJS
REST APIs

Data & Tools

Pandas
NumPy
Power BI
Tableau
Git
GitHub
Jupyter Notebook
Hugging Face
FFmpeg
07.

Certifications Vault

DeepLearning.AI Nov 2024

Generative AI with Large Language Models

ID: VR3ZL6II1KGI

Skills Gained

Generative AI Tools, Large Language Modeling, RAG, Prompt Engineering, Fine-Tuning

University of California, Davis Jul 2024

SQL for Data Science

ID: YRRTQUFDKK99

Skills Gained

Data Analysis, SQLite, SQL, Data Formatting

DataCamp Aug 2024

Data Analyst Associate

ID: DAA0019119697492

Skills Gained

Data Visualization, Analytical Skills, Data Cleaning, Exploratory Analysis

HackerRank Aug 2024

SQL (Intermediate) Certificate

ID: 72d70c5a4110

Skills Gained

Problem Solving, SQL Query Optimization, Complex Joins, Subqueries

Forage Aug 2024

BCG - Data Science Job Simulation

ID: EssdLw7NEqtrtrtYY

Skills Gained

Time Management, Linear Regression, Feature Engineering, Business Strategy, Modeling

University of London Jul 2024

Probability and Statistics: To p or not to p?

ID: PVZPKYSUFMLV

Skills Gained

Applied Mathematics, Probability, Hypothesis Testing, Statistical Analysis

IBM Jul 2024

Generative AI: Elevate Your Data Science Career

ID: 5XVP7RT26RNU

Skills Gained

Fine Tuning, Large Language Models (LLM), Generative AI Productivity

John Hopkins University Jul 2024

Business Analytics with Excel: Elementary to Advanced

ID: ML5P6DCEASQF

Skills Gained

Business Decision Making, Spreadsheets, Excel Modeling, Decision Analytics

Google Jul 2024

Google Data Analytics Professional Certificate

ID: FRYF2NHRR7BG

Skills Gained

Data Cleaning, SQL, R Programming, Data Visualization, Problem Solving

University at Buffalo Jun 2024

Blockchain Basics

ID: 8fab764bb040384b5b3a6e40a14df96b

Skills Gained

Algorithms, Problem Solving, Distributed Ledgers, Cryptography Basics

Udemy Aug 2022

Python for Data Science and Machine Learning Bootcamp

ID: UC-b1f5f0e4-f8e5-44e3-b3e5-af83f2931ced

Skills Gained

Python (Programming Language), Data Science, Pandas, NumPy, Scikit-Learn

Altair RapidMiner Jun 2024

Machine Learning Professional Certificate

ID: RapidMiner-9892

Skills Gained

Fine Tuning, Large Language Models (LLM), Predictive Analytics, Modeling

08.

Education

Bachelor of Technology (B.Tech) – Computer Science (Data Science)

Sharda University, Greater Noida

Specialized curriculum in advanced statistical models, deep neural networks, computer vision grids, and database query architecture.

2021 – 2025 CGPA: 8.08 / 10
09.

Visual Playground

Step into the sandbox. Tap any bubble in the interactive physics field to experience its generative workflow. Floating nodes represent active models, RAG vector spaces, and in-browser utilities.

CATEGORY Node Tooltip

Bubble Title

Detailed description of the tool or project simulator goes here.

Click to Launch
10.

Contact Terminal

Direct Connections

Reach out directly via secure voice/mail channels or review full codebases on GitHub.

Response Latency: < 12h Active Node
mail_gateway.exe

// Establish secure communication bridge

>
>
>
Status: Connected
PORTFOLIO — SAHIL.SINGH

Design

COMPILING CORE DATASETS 000
About Sahil 01. Flagship Product 02. Work Experience 03. Selected Works 04. Publications 05. Metrics & Achievements 06. Technical Expertise 07. Certifications Vault 08. Education 09. Visual Playground 10. Contact Terminal
Technology Stack Trace Integration