AI/ML Engineer | Building Production Systems | Learning in Public
I'm a 4th-year AI & ML student at GGSIPU, New Delhi turning ideas into production code. I don't just implement algorithms—I build end-to-end systems that actually work, deploy reliably, and solve real problems.
Currently focused on ML engineering (production systems, MLOps, deployment), LLM applications (RAG, vector databases), and financial ML (trading systems).
- Education: B.Tech AI & ML | GGSIPU, New Delhi (2023–2027)
- CGPA: 7.5/10
- Status: 4th year, actively seeking full-time roles, internships, remote opportunities, freelance projects
- Specialization: ML Engineering (from notebook to production)
- Location: New Delhi, India (open to remote)
- Hackathon Finalist — JECRC Innov8
- Smart India Hackathon Participant — National-level competition
- IITM Matrix 2.0 Participant — ML/AI track
- FPGA Design Flow Workshop — NIT Delhi
- Data Privacy & Confidentiality Program — NIT Delhi + IIT Roorkee
- Certifications: Microsoft Azure Fundamentals (AZ-900), Python for AI & Development, Java Core
Vision Price Net — Multimodal ML System
Live Demo: https://house-prediction-model-1rtp.onrender.com/
House price prediction combining Deep Learning (ResNet18) + XGBoost with tabular data.
Results:
- ✅ 88% R² Score (explains 88% of price variance)
- ✅ 2,500–5,000 house dataset
- ✅ <2.5 second inference time
- ✅ Live on Render (fully deployed)
Tech: PyTorch, TensorFlow, XGBoost, Flask, Render
Why it matters: Shows I can handle real-world complexity—multiple data types, ensemble methods, model comparison, deployment decisions.
Binance Futures Trading Bot — REST API Backend
Production-grade trading bot demonstrating REST API design, HMAC-SHA256 authentication, and secure backend development.
What I built:
- ✅ Binance Futures Testnet integration (live order execution)
- ✅ MARKET & LIMIT order functionality
- ✅ HMAC-SHA256 signature generation
- ✅ Modular, testable Python architecture
- ✅ Structured logging and error handling
Tech: Python, REST APIs, HMAC authentication, argparse, structured logging
Why it matters: Shows I understand API security, credential handling, and production backend practices—not just ML algorithms.
MLOps Batch Signal Pipeline — Production Infrastructure
Real-world batch pipeline showing how professional ML teams build systems—not just models.
What I built:
- ✅ Deterministic batch processing (reproducible execution)
- ✅ YAML configuration management
- ✅ Docker containerization
- ✅ Structured logging & JSON metrics
- ✅ Production observability practices
Results:
- ✅ Processes 10,000+ rows reliably
- ✅ 100% reproducible (seeded randomness)
- ✅ Runs identically in Docker & local
Tech: Docker, pandas, NumPy, YAML, structured logging
Why it matters: The hard part of ML isn't the algorithm—it's the infrastructure. This project shows I understand deployment, reproducibility, and observability.
AI-Powered Payroll System — Microservices Architecture
Distributed system demonstrating system design, microservices communication, and fault tolerance.
What I built:
- ✅ Java backend (payroll calculations, data persistence)
- ✅ Python FastAPI microservice (Isolation Forest anomaly detection)
- ✅ Real-time ML inference in business workflow
- ✅ Graceful degradation (works even if ML service is down)
- ✅ REST API communication between Java & Python
Tech: Java 8, FastAPI, Scikit-learn, CSV persistence, HTTP APIs
Why it matters: Shows I can design systems where business logic and AI work together reliably—a real production concern.
Python (expert), Java, SQL
- Model Building: Scikit-learn, XGBoost, TensorFlow, PyTorch
- Techniques: Supervised & unsupervised learning, ensemble methods, anomaly detection (Isolation Forest)
- Feature Engineering: pandas, NumPy, data cleaning, feature selection
- Evaluation: Cross-validation, model comparison, performance metrics
- Visualization: matplotlib, seaborn
- LangChain, ChromaDB, RAG pipelines
- OpenAI & Anthropic APIs
- Vector databases, prompt engineering
- Containerization: Docker (reproducibility, deployment)
- CI/CD: GitHub Actions (automated testing, deployment)
- ML Tracking: MLflow (experiment tracking)
- Testing: pytest (test-driven development)
- Logging: Structured logging, JSON metrics, observability
- Configuration: YAML, environment variables
- Frameworks: FastAPI, Flask (REST API development)
- Databases: MySQL, CSV persistence
- Security: HMAC-SHA256 authentication, API key management
- Deployment: Gunicorn, Render (PaaS), Uvicorn
- Exploration: Microsoft Azure
- Languages: HTML5, CSS3, JavaScript
- UI/UX: Responsive design, user feedback (animations, loading states)
- Development: Jupyter, VS Code, Git/GitHub
- CLI Tools: argparse, python-dotenv
- Deployment: Render (PaaS)
LLM Applications & RAG Pipelines
- Building question-answering systems with LangChain
- Exploring vector databases (ChromaDB) for semantic search
- Working with OpenAI & Anthropic APIs
- Learning prompt engineering best practices
Why? This is where AI is moving—from static models to dynamic, context-aware systems.
- AI & Machine Learning Algorithms
- Database Management Systems (DBMS)
- Data Structures & Algorithms
- Web Development & APIs
I'm interested in:
- ✅ Full-time ML Engineering roles (production systems, MLOps)
- ✅ Internships (summer 2026, 6+ months)
- ✅ Remote opportunities (anywhere in India)
- ✅ Freelance projects (specific ML/data tasks)
- ✅ Collaboration (interesting problems, learning opportunities)
I'm NOT interested in:
- ❌ Cookie-cutter projects with no learning
- ❌ Resume-padding exercises
- ❌ Work that's purely theoretical (I like building real things)
On Building: Code should be:
- ✅ Working (deployed, not just in a notebook)
- ✅ Reliable (error handling, edge cases, logging)
- ✅ Maintainable (clean, modular, documented)
- ✅ Learnable (reflects current best practices)
On Learning: I learn by:
- ✅ Building real projects (not tutorials)
- ✅ Shipping code to production
- ✅ Learning from failures (bugs, rollbacks, lessons)
- ✅ Reading production code (open source, papers)
On Hiring Managers: I want to work with teams that:
- ✅ Value learning and growth
- ✅ Ship real products (not toys)
- ✅ Care about code quality (not just velocity)
- ✅ Invest in junior engineers
Email: himanshubg70@gmail.com
LinkedIn: https://www.linkedin.com/in/himanshu-kumar-076a13321/
GitHub: https://github.com/Himan-stack
Got a project idea? Want to collaborate? Have feedback?
I'm always open to interesting conversations about ML, system design, or building cool stuff.
Real projects. Real code. Real deployment.
Not a résumé builder. A learning engineer building in public.
Last updated: June 2026