AI Resume Screening
An AI-powered resume analysis and ranking system combining semantic similarity with LLM-based information extraction.
Hybrid semantic + structured scoring approach.
- Python
- FastAPI
- React
- Sentence Transformers
- MiniLM
- +3
Utkarsh Srivastava
AI/ML Engineer|Computer Science Student
Building intelligent systems through machine learning, deep learning, and modern AI engineering.
I'm a B.Tech Computer Science student focused on machine learning, deep learning and generative AI. My work spans model training and evaluation, retrieval-augmented generation and LLM applications, computer vision, and the backend/API layer that turns a model into a usable product.
AI/ML
Machine Learning & Deep Learning
Projects
Multiple End-to-End AI Systems
Focus
AI Engineering & Generative AI
Education
B.Tech Computer Science
01 / About
About Me
I'm a B.Tech Computer Science student specializing in AI/ML, interested in building practical intelligent systems rather than notebooks that never leave the laptop.
My focus areas are machine learning, deep learning, NLP, computer vision, LLMs and RAG — paired with the AI engineering and backend work needed to serve them: FastAPI services, clean data pipelines, and honest evaluation.
Currently exploring
- LangChain
- LangGraph
- RAG systems
- LLM evaluation
- AI agents
02 / Projects
Featured Projects
End-to-end systems — each one has a problem, an architecture and a working implementation. Open a case study for the full pipeline.
VisionSpeak
A computer vision and image understanding system that turns what the camera sees into spoken natural language.
Vision-to-language pipeline built for real-time use.
- Python
- OpenCV
- PyTorch
- Vision Models
- Text-to-Speech
Image Captioning
An encoder–decoder captioning model: CNN image encoder, LSTM decoder with an attention mechanism, trained on Flickr8k.
CNN encoder + LSTM decoder with attention, BLEU-evaluated.
- Python
- PyTorch
- CNN
- LSTM
- Attention
- +2
CityLifeScore
A livability scoring system for Indian cities that combines six weighted dimensions into a single comparable score.
Explainable, dimension-level scoring instead of a single opaque number.
- Python
- Pandas
- Scikit-learn
- Data Visualisation
Driver Drowsiness Detection
A real-time computer vision system that detects driver drowsiness from a webcam feed and raises an alert.
Real-time detection with a temporal window to suppress false alarms.
- Python
- OpenCV
- TensorFlow
- CNN
- Real-Time Inference
03 / Skills
Technical Skills
Tools and concepts I actively use across modelling, generative AI, vision and the backend layer.
Programming
- Python
- Java
- C
- SQL
Machine Learning
- Scikit-learn
- Feature Engineering
- Model Evaluation
- Supervised Learning
- Unsupervised Learning
Deep Learning
- PyTorch
- TensorFlow
- CNNs
- RNNs
- LSTMs
- Attention Mechanisms
Generative AI
- LLMs
- RAG
- Embeddings
- Vector Databases
- Prompt Engineering
- LangChain
- LangGraph
Computer Vision
- OpenCV
- Image Classification
- Image Captioning
- Object Detection
Backend / Development
- FastAPI
- REST APIs
- React
- Git
- GitHub
04 / Journey
Learning Journey
A progression rather than a job history — how the computer science foundation turned into AI engineering work.
Foundation
B.Tech Computer Science — DIT University
Core computer science coursework building the systems foundation everything else sits on.
- Data Structures & Algorithms
- DBMS
- Operating Systems
- Computer Networks
- Computer Organization
- Compiler Design
- Theory of Computation
Step 01
Machine Learning
Supervised and unsupervised learning, feature engineering, and honest model evaluation with Scikit-learn.
- Scikit-learn
- Feature Engineering
- Model Evaluation
Step 02
Deep Learning
Neural network architectures in PyTorch and TensorFlow — CNNs for vision, RNNs/LSTMs and attention for sequences.
- PyTorch
- TensorFlow
- CNNs
- LSTMs
- Attention
Step 03
Generative AI
LLM applications, embeddings and vector search, retrieval-augmented generation, and prompt design.
- LLMs
- RAG
- Embeddings
- Vector Databases
Now
AI Engineering
Shipping models as services: FastAPI backends, evaluation loops, agent workflows with LangChain and LangGraph.
- FastAPI
- LangChain
- LangGraph
- LLM Evaluation
05 / Achievements
Achievements
ML Hackathon — Runner-Up
[EVENT NAME] · [YEAR]
Runner-up placement in a machine learning hackathon.
Hackathon Participation
[EVENT NAMES]
Built and presented working prototypes under time constraints.
ML Competitions
[PLATFORM / COMPETITION]
Participated in applied machine learning competitions.
Coding Practice
LeetCode
Consistent data structures and algorithms practice.
Academic
[ACHIEVEMENT]
[Add a relevant academic achievement here.]
06 / Developer
Developer Profile
Values below are configured manually in the content file — no live statistics are fabricated.
GitHub
View profile →- Public repositories
- 7
- Primary languages
- Python · Java
- Contribution activity
- [UPDATE MANUALLY]
LeetCode
View profile →- Problems solved
- 600+
- Rating
- 2109
- Focus
- DSA · Problem Solving
07 / Education
Education
DIT University
B.Tech, Computer Science & Engineering
AI/ML specialization
Relevant coursework
- Data Structures & Algorithms
- DBMS
- Operating Systems
- Computer Networks
- Computer Organization
- Compiler Design
- Theory of Computation
- Machine Learning
Want the complete picture?
View or download my resume for a detailed overview of my projects, technical skills, and experience.
08 / Contact
Let's Build Something Intelligent
Open to internships, AI/ML roles and collaboration on applied machine learning projects.