Sahil Bhatti — AI Backend Engineer

selected projects

Systems I'm building.

01 AI · Systems Engineering In Progress

The Lazarus Engine

Code-analysis prototype for understanding, rewriting, and reviewing legacy modules with an AI-assisted backend workflow.

Problem

Legacy modernisation fails when teams treat old code as plain text. Lazarus Engine explores a safer backend pipeline: parse the source, map behaviour and dependencies, generate a modern target draft, synthesize tests, and keep a human review step before any migration decision.

Architecture

# Lazarus Engine — pipeline overview LegacyParser → AST extraction # tree-sitter SemanticMapper → intent graph # LLM reasoning DepResolver → dep tree # static analysis CodeGenerator → modern target # draft rewrite TestSynthesizer → pytest suite # auto-generated DiffReviewer → human review # review bundle

⌥ GitHub not public yet → Case study pending

Current proof

Build

In progress

Private

Repository

Review

Human checkpoint

Tests

Verification focus

Stack

Python Claude API tree-sitter FastAPI PostgreSQL Docker Pydantic v2 pytest

02 RAG · Backend API Prototype

Multi-tenant RAG Chatbot Engine

Backend project for document ingestion, vector search, grounded answers, and lead capture workflows.

Problem

Content-heavy businesses need answers from their own documents without hallucinated claims. This project focuses on the backend pieces: ingestion, chunking, embeddings, tenant-aware storage, source retrieval, Claude-powered responses, and a simple lead capture path.

Architecture

# Ingestion pipeline Crawler → FireCrawl # HTML/PDF/DOCX Chunker → text splits # overlap windows Embedder → vectors # embedding model pgvector → Supabase # vector storage # Query pipeline Query → embed → retrieve # source search Context → Claude # grounded answer Response → API stream # frontend widget

⌥ GitHub not public yet → Demo not published yet

Current proof

RAG

Core pattern

Tenant

Isolation model

Sources

Grounding focus

Leads

Capture flow

Stack

FastAPI pgvector Claude API Supabase Docker Pydantic v2 SSE

03 ML · Security · Research M.Tech Research

Encrypted Network Traffic Classifier

Academic ML research on classifying encrypted traffic using flow-level statistics without decrypting payloads.

Problem

Deep packet inspection becomes limited when traffic is encrypted. This M.Tech research project studied whether statistical flow features such as inter-arrival times, packet length distributions, and burst patterns can support traffic classification without inspecting payload contents.

Approach

# Feature engineering pipeline Capture → tshark # PCAP → flows Features → flow stats # IAT, pkt_len, burst Model → ML compare # classical ML Validate → report # academic evaluation

⌥ GitHub not public yet → PDF not attached

Current proof

M.Tech

Research work

PCAP

Traffic data

Stats

Flow features

No DPI

Payload privacy

Stack

Python XGBoost scikit-learn tshark pandas PCAP

04 ML · Medical Imaging Research

COVID-19 Chest X-Ray Classifier

Medical-imaging ML project using transfer learning and a FastAPI-style inference interface.

Problem

Built a learning-focused chest X-ray classification pipeline around transfer learning, augmentation, model evaluation, and API-style serving. The project is positioned as ML engineering practice, not a medical product.

Pipeline

Dataset → CXR images # train/valid split Augment → transforms # image variance Model → CNN # transfer learning Evaluate → metrics # validation report Serve → FastAPI # inference API

⌥ GitHub not public yet

Current proof

Model pipeline

API

Serving pattern

Image task

Practice

Learning project

Stack

PyTorch ResNet-50 FastAPI Albumentations CUDA Docker

experience & education

Where I've worked and studied.

2026 — Present

AI Backend Engineering Projects

bysahil.dev · Portfolio Work

Building portfolio-grade AI backend systems around RAG, FastAPI, pgvector, LLM APIs, and verification-heavy backend workflows while preparing for a first full-time AI Backend Engineer role.

Claude API FastAPI pgvector RAG

2020 — 2022

M.Tech in Data Science

Indian Institute of Technology, Jammu

Thesis: Encrypted network traffic classification using flow-level statistical features. Coursework in ML, deep learning, distributed systems, statistical inference.

ML/DL Research Python Statistics

2015 — 2019

B.Tech in Computer Science

Guru Nanak Dev University, Amritsar

Core CS foundations: data structures, algorithms, DBMS, OOP, OS, networks. Industrial training at Software Cell, GNDU.

C/C++ Java DBMS Algorithms

about

The person
behind the commits.

I'm Sahil Bhatti — 28, based in Punjab, India. I build AI backend systems with a verification-first mindset: every claim should be grounded in explicit knowledge, logic, or stated uncertainty.

My technical instincts come from a decade of working close to the metal — C/C++ before Python, algorithms before frameworks. That foundation shows in the systems I design: low-latency, composable, easy to debug when they inevitably break.

Currently building The Lazarus Engine, a RAG chatbot backend, and ML/API projects to show practical backend engineering depth: data modelling, API design, validation, testing, deployment readiness, and clear technical communication.

When I'm not writing code: reading hard sci-fi and non-fiction (currently: Project Hail Mary), riding my Royal Enfield Classic 350, swimming, and playing guitar.

🏛️

M.Tech · Data Science

IIT Jammu

2020 – 2022

🎓

B.Tech · Computer Science

GNDU Amritsar

2015 – 2019

Currently reading

Project Hail Mary — Andy Weir Hard Sci-fi

Designing Data-Intensive Applications Systems

The Pragmatic Programmer Engineering

contact

Let's build something
that actually works.

Open to full-time AI Backend Engineer and Backend Engineer roles. Remote is preferred, but I'm open to the right on-site or hybrid opportunity. Response within 24 hours.

✉️

sahil@bysahil.dev

→ 💼

linkedin.com/in/sahilarchitect

→ ⌥

GitHub

github.com/SahilArchitect

→

sahil@portfolio ~ candidate_profile.json

$ cat candidate_profile.json
{
  "name": "Sahil Bhatti",
  "role_target": "AI Backend Engineer",
  "education": "M.Tech · IIT Jammu",
  "stack": [
    "Python", "FastAPI", "Claude API",
    "pgvector", "Docker", "C/C++"
  ],
  "strengths": [
    "RAG system design",
    "Legacy modernisation",
    "Low-latency inference pipelines"
  ],
  "available": true,
  "full_time": true,
  "remote_preferred": true,
  "response_sla": "<24h"
}

$ echo $STATUS
→ Open to opportunities. Let's talk.
$

Systems I'm building.

Where I've worked and studied.

The personbehind the commits.

The person
behind the commits.