open to ML / backend / full-stack roles · 2026

I build systems thatactually ship.

I'm Tanisha Agarwal — an ML engineer and full-stack developer working across computer vision, backend architecture, and DevOps for fintech, banking, and insurance.

ml-engineerfull-stack-devcomputer-vision
tanisha@portfolio — zsh
~whoami
ML engineer · backend developer · AI enthusiast
~cat focus.txt
computer vision · distributed systems · AWS · Kubernetes
~echo $motto
turning ideas into intelligent, scalable software
~status --now
● open to ML / backend / full-stack roles ▸ Mumbai · IST
~
pipeline: 90s → 4saccuracy: 100%status: 200 OK
ReactNext.jsTypeScriptTailwind CSSNode.jsPythonPostgreSQLASP.NET Corescikit-learnC++GitFramer MotionReactNext.jsTypeScriptTailwind CSSNode.jsPythonPostgreSQLASP.NET Corescikit-learnC++GitFramer Motion
01about

I ship ML and full-stack systems that survive real users.

I work end to end — from computer-vision models and data pipelines to the backends, deployments, and monitoring around them — with a bias for reliability in regulated environments.

01

Computer vision

Production detection, classification, and OCR pipelines built for accuracy in live financial workflows.

YOLOFaster R-CNNOpenCVPyTesseract
02

Applied ML

Feature engineering, cross-validation, and tuning on structured data — models chosen by precision, recall, and F1.

PyTorchscikit-learnXGBoostPandas
03

Full-stack engineering

Java / Spring Boot and Node.js backends with JWT auth and REST APIs, wired to modern, responsive frontends.

Spring BootNode.jsRESTMongoDB
04

DevOps & observability

Containerized deployments instrumented with metrics and tracing so failures are visible before users feel them.

DockerKubernetesCI/CDAWSPrometheusJaeger
05

Compliance-aware ML

Systems built for regulated finance: PII masking, auditable rule logic, and documented data-quality standards.

DPDP / DPDPAPII maskingaudit logs
06

Cross-functional delivery

Translating messy requirements for non-technical stakeholders across banking, insurance, and product teams.

Agile / ScrumJirastakeholder comms
throughput

0×

faster document processing

90s → 4s per document

accuracy

0%

processing accuracy in live onboarding

YOLO · Faster R-CNN · 3+ doc types

automation

0%+

manual review effort eliminated

OCR PII masking · DPDP-compliant

performance

0%

backend latency cut at NepByte

server-side caching + response tuning

02selected work

Systems built to run in production, not demos.

A mix of computer vision, applied ML, and full-stack builds — each measured by what it did once real users and real data hit it.

01 / 03 · 2026

Computer Vision · Production

Document Intelligence Pipeline

End-to-end detection and classification across 3+ document types for live financial onboarding, with an OCR-based PII masking layer for DPDP compliance.

YOLOFaster R-CNNPyTesseractOpenCVDocker
impact90s → 4s · 100% accuracy
02 / 03 · 2025

Machine Learning

IPL Match Outcome Predictor

A cross-validated ML pipeline with feature engineering and hyperparameter tuning on large structured datasets, with the model chosen by precision, recall, and F1.

Pythonscikit-learnXGBoostPandas
impactData-driven model selection
03 / 03 · 2025

Full-Stack · Real-Time

Treasure Hunt Platform

A multi-user system deployed at a live college event with real-time state, a MongoDB-indexed backend, and JWT auth — handling concurrent sessions at 100% uptime.

Node.jsExpress.jsMongoDBJWT
impact100% uptime under load
03experience

Five internships, one throughline — ship what actually works.

Across insurance, banking, and product teams — plus the coursework behind it.

  1. 01

    computer vision

    Product & ML Engineer Intern

    Star Union Dai-ichi Life Insurance · Mumbai

    Jan 2026 — Jun 2026

    Built an end-to-end document pipeline (YOLO, Faster R-CNN) for automated detection and classification across 3+ document types, containerized with Docker and wired into live onboarding — cutting turnaround from 90s to 4s at 100% accuracy. Added an OCR-based PII masking system for DPDP compliance, removing 80%+ of manual review.

    YOLOFaster R-CNNOpenCVPyTesseractDocker
  2. 02

    ai research

    Research Intern — AI in Banking

    HDFC Bank · Mumbai

    Jun 2025 — Jul 2025

    Analyzed AI/ML adoption across 5+ global banks and mapped scalable automation use cases to compliance and operational workflows. Evaluated rule-based vs. model-driven approaches for transaction monitoring and fraud detection, identifying standardization opportunities across portfolios.

    ML StrategyFraud DetectionTransaction Monitoring
  3. 03

    backend · full-stack

    Backend & Frontend Developer

    NepByte Technologies · Vellore

    Sep 2024 — May 2025

    Led a 4-member team delivering production platforms end to end: RESTful backend with JWT auth, structured logging, and CI/CD. Drove a 30% performance improvement via server-side caching, and added end-to-end observability with Prometheus metrics and Jaeger tracing across the microservices stack.

    Spring BootREST · JWTCI/CDPrometheusJaeger
  4. 04

    process automation

    Summer Technology Intern

    SBI Mutual Fund · Mumbai

    Jun 2024 — Jul 2024

    Overhauled an internal operations platform serving 50+ stakeholders, reducing task completion time by 25% and removing manual bottlenecks through rule-based order-request automation.

    Rule EnginesSQLWorkflow Automation
  5. 05

    frontend

    Web Development Intern

    The Times Group · Mumbai

    Jul 2023

    Modernized a legacy JSP platform with responsive layouts, cutting page latency by 33% (900ms → 600ms) and hardening role-based access controls.

    JSPResponsive UIRBAC

Education

— 2022–2026

B.Tech, Computer Science & Engineering

2022 — 2026

Vellore Institute of Technology

GPA 8.53 / 10

National semi-finalist at Flipkart GRiD 7.0 (Engineering Track) and Nokia “Accelerate Her in Tech”. Creative Head of the ADMARK Club and a Hacktoberfest open-source contributor.

Relevant coursework

Data Structures and AlgorithmsDesign and Analysis of AlgorithmsObject-Oriented ProgrammingDatabase Management SystemsOperating SystemsComputer NetworksTheory of ComputationArtificial IntelligenceMachine LearningNatural Language ProcessingComputer VisionSoftware Engineering

Higher Secondary (Class XI–XII), Science

2020 — 2022

Poddar Brio International School

PCM + Computer Science

Physics, Chemistry, and Mathematics with Computer Science — early grounding in programming and quantitative problem-solving.

Core stack

19 tools
JAVAJavaPYPythonC++C++JSJavaScriptPTPyTorchSKscikit-learnXGXGBoostYOLOYOLORCNNFaster R-CNNCVOpenCVSBSpring BootNODENode.jsEXExpress.jsAPIREST APIsDBSQL / MongoDBAWSAWSDOCDockerCICI/CDGITGit
04contact

Let's build something that actually matters.

available for work

Get in touch

I'm open to early-stage ML roles, backend positions, and full-stack opportunities in fintech and B2B tech — and always up for a good engineering conversation.

Email me directly
TA

Send a message

goes straight to my inbox

online

I'll only use your email to reply — never shared.