Senior Full Stack Engineer → AI/ML Engineer
Building intelligent systems at the intersection of backend engineering and applied AI.
10+ years shipping production systems · Currently focused on AI/ML engineering
RAG system for querying multi-year company annual reports
Extracts text, charts, and tables from PDFs. Stores embeddings in Pinecone, structured data in metadata. GPT-4o Vision describes visual pages so charts become searchable. Supports cross-year queries and auto-generates matplotlib charts from table data.
PythonOpenAI GPT-4oPineconePyMuPDFpdfplumberRAGVector Search
AI-powered mock interview system for ML/System Design/Observability roles
Full mock interview flow — pick a question semantically via Pinecone, submit your answer, get scored by GPT-4o with strengths/gaps/follow-up questions, and receive a final readiness report. 30 seed questions across ML theory, system design, and ML observability.
FastAPIOpenAI GPT-4oPineconeElasticsearchPythonRAGVector Search
Elasticsearch handles structured filters (topic, difficulty, tags) and stores ideal answers.
Pinecone handles semantic question similarity search.
GPT-4o evaluates answers and generates follow-up questions dynamically.
High-performance microservice built in Rust
Exploring systems programming with Rust — focused on memory safety, async performance, and building production-grade backend services without a garbage collector.
RustAsyncSystems Programming
LAN device discovery tool in Rust
Scans local network using the r_lanlib crate to detect connected devices. Built during Rust learning phase to explore low-level networking in a systems language.
RustNetworkingr_lanlib
Working through Rust from first principles
Structured notes and exercises covering ownership, borrowing, lifetimes, traits, async/await, and error handling — the concepts that matter for building reliable backend systems.
RustSystems Programming
Python-based stock analysis experiments
Quantitative analysis scripts using Python — exploring financial data pipelines, technical indicators, and data visualization. Foundation for future ML-based trading signal work.
PythonJupyterPandasFinancial Data
Zero to Mastery deep learning curriculum
Hands-on notebooks covering CNNs, transfer learning, NLP with transformers, and time-series forecasting using TensorFlow/Keras.
TensorFlowKerasDeep LearningPythonJupyter
Production ML Engineer
├── RAG systems & vector search
├── LLM fine-tuning & evaluation
├── ML observability & monitoring ├── Real-time feature stores
└── High-performance backends in Rust
I spent 10+ years building scalable backend systems at companies like Shaadi.com, Sourcebits, and Wheelstreet. Now applying that production engineering discipline to AI/ML — the difference between a demo and something that works in prod.
| Title | Topic | |
|---|---|---|
| 🧠 | Beyond RLHF: Moving to RLAIF — The AI Feedback Loop | LLM Alignment · RLAIF vs RLHF · Constitutional AI |
| ⚡ | Why Every AI/ML Engineer Must Master Quantization in 2026 | Model Compression · INT8 · QLoRA · Inference Optimization |
| 🏗️ | Understanding the Architecture of Modern LLMs — Layer by Layer | Transformer Architecture · Attention · MoE · KV Cache |
| 🔍 | Advanced RAG Preprocessing: How Top AI Teams Are Making Retrieval Smarter | RAG · Chunking Strategies · Hybrid Search · Reranking |
- Vue.js State Management with Vuex — Medium
- Writing Clean Code Principles — Medium
- Why Dashboards Are Valuable — LinkedIn
Open to Senior Backend / AI/ML Engineering roles · Pune, India · Remote-friendly