Shweta Pasi
AI Engineer
Management: 1-2 years (1-5 people)
Python
SQL
TypeScript
JavaScript
Active 8 days ago
Download CVCV
Overview
Technical skills
Projects
Timeline
Roles
Overview
Backend engineer (mid-level) with a focus on data pipelines and client-side data engineering; strongest at shipping deterministic pre-aggregation pipelines and compact client encodings.
Phone
Technical skills
Languages
4
Python
SQL
TypeScript
JavaScript
AI/ML
10
NumPy
Pandas
LLM
ChatGPT
Claude
Gemini
LangChain
Scikit-learn
Machine Learning
Spark
Frontend
6
React.js
Next.js
Tailwind CSS
Cytoscape.js
Element Plus
Recharts
DevOps
5
AWS
Azure
Amazon S3
GitHub
Vercel
Other
17
PostgreSQL
Power BI
FastAPI
Tableau
Canva
Docker
Git
Rest API
Supabase
pySpark
Groq
AI Agents
RAG
Embeddings
Knowledge Graph
Multi-Agent Systems
Semantic Search
Projects
Created a CSV-to-interactive knowledge graph pipeline that turns 5,000+ rows into connected graph visualizations. The app includes relationship-confidence scoring and row-level provenance to support traceability. The backend logic is validated with backend tests and the UI uses Cytoscape.js.
Built a multi-tenant document generation SaaS that creates multiple enterprise document types using Gemini-based AI drafting. The platform supports tenant isolation, version control, and DOCX export with JWT/OAuth authentication. Implemented with React, FastAPI, and MongoDB.
Shipped a free-first AI tool recommendation engine that ranks 67+ tools across categories using real pricing and usage limits. The system prioritizes verified free plans while aiming to keep infrastructure costs at zero. Built as a Next.js app with TypeScript, Supabase, and Groq integration.
Built a diagnostics console to analyze LLM agent execution traces and flag issues such as redundant tool calls, context bloat, model overuse, and retry loops. The tool estimates projected savings per diagnosis and provides cost forecasting for agent runs. Implemented as a web app with a FastAPI backend and React frontend.
Timeline
Assistant Manager – Data Analyst
·
Lead
CEAT Tyres (RPG Enterprises)
·
Full-Time
Built and deployed the organization’s first production GenAI budgeting application with LLM-driven expense categorization and natural-language summaries using FastAPI/PostgreSQL and React/TypeScript. Engineered industrial analytics across multiple Ignition SCADA workstreams, including OEE monitoring, predictive maintenance, and energy savings. Digitized 300+ paper quality/maintenance checklists into sensor-driven workflows and delivered Tableau dashboards and AWS QuickSight reporting for leadership stakeholders.
FastAPI
PostgreSQL
React.js
TypeScript
Tableau
AWS
D.Y. Patil University
Master's Degree ·
Artificial Intelligence & Machine Learning
Graduate Engineer Trainee
·
Junior
CEAT Tyres (RPG Enterprises)
·
Full-Time
Developed a Python automation pipeline for export containerization planning using Pandas and NumPy, reducing the planning cycle time and eliminating recurring spreadsheet errors. Performed Python/SQL analytics (including RFM modeling and bottleneck identification) and created executive dashboards in Tableau and Power BI for multiple business KPIs such as policy goals, warranty claims, rubber analytics, attrition, and headcount.
Python
Pandas
NumPy
SQL
Tableau
Power BI
Junior Backend Developer
Confidence: Medium Data Platform
Backend engineer (mid-level) with a focus on data pipelines and client-side data engineering; strongest at shipping deterministic pre-aggregation pipelines and compact client encodings.
API Design
How well APIs are designed
Not evidenced in public code
Data Layer & Database
Working with databases
Not evidenced in public code
Scalability & Performance
Handling load and speed
Not evidenced in public code
System Architecture
Overall system structure
Not evidenced in public code
Security & Auth
Protecting data and access
Not evidenced in public code
Reliability & Observability
Stability and monitoring
Not evidenced in public code
Expertise
Python· Junior
Industries
Lifestyle· Junior
Transportation & Logistics· Junior
Technologies
Rest API
FastAPI· since 2025
Recommendations
- Build end-to-end ETL and pre-aggregation pipelines that emit compact, verified binary artefacts for low-latency dashboards.
- Implement reproducible verification and CI checks that assert parity between server-side and client-side aggregates (like the existing verify pattern).
- Productize anomaly detection by adding production inference pipelines, monitoring and safe retraining workflows.
- Expose a small read API (with pagination, idempotency and authenticated access) to serve precomputed aggregates for integration with other services.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Junior Frontend Developer
Confidence: Low Fullstack
Fullstack frontend engineer (Middle) specializing in system-first web architecture and pragmatic performance trade-offs. The strongest proven skill is systems and product architecture as documented in askforai-architecture/index.html, which details server components, prerendering, an LLM adapter, and a resilience model. What is not evidenced in the human-authored files is production component implementation, robust frontend async state logic (cancellation, optimistic updates, state machines) or meaningful test coverage.
UI Component Architecture
4/10
How interface parts are built
Architectural-first UI decisions are documented (Server Components, islands of interactivity, single-file editorial weights) but no human-authored component library or non-trivial component code is present to prove component-level design and composition choices.
Evidence
askforai-architecture/index.html: 'Browser — React Server Components + islands of interactivity'
askforai-architecture/index.html: 'The weights live in a single file so the product\'s editorial stance is one reviewable diff'
Responsive & Cross-browser
5/10
Works on all screens and browsers
Responsive and cross-device concerns are explicitly considered (viewport, prerendering, LCP budget, CSS variables and dark-mode guidance), and the human-authored doc includes concrete responsive CSS patterns inlined in the static HTML.
Evidence
askforai-architecture/index.html: meta viewport and references to LCP budget '<2s'
askforai-architecture/index.html: inline CSS with @media and prefers-color-scheme rules
Performance Optimization
5/10
Speed of the interface
Clear, measurable performance-first architecture is documented (server-first rendering, prerendered pages, minimal client JS islands, LCP budget), but there is no human-authored performance benchmark code or before/after artifacts in the human-authored files.
Evidence
askforai-architecture/index.html: 'Server-first rendering' and 'Client JavaScript is limited to genuinely interactive islands'
askforai-architecture/index.html: 'LCP budget <2s' and 'Prerendered surface'
Accessibility & Semantics
5/10
Usable for everyone
Accessibility and semantics are considered in the human-authored document and the static HTML includes ARIA roles and descriptive attributes, but there is no test automation or axe/CI evidence in the human-authored files.
Evidence
askforai-architecture/index.html: SVGs with role='img' and aria-labels in diagrams
askforai-architecture/index.html: use of aria-label attributes on navigation and sections
State Management & Data Flow
4/10
Managing data in the app
Solid server-state and data-flow architecture is described (data-access module with Postgres fallback, deterministic ranking, batched freshness jobs), but human-authored files do not contain implementation of request cancellation, optimistic updates, or state machines to verify discipline in code.
Evidence
askforai-architecture/index.html: 'All reads pass through one data-access module that tries Postgres and falls back to the compiled seed'
askforai-architecture/index.html: 'One request flows through six discrete stages' and 'Deterministic core'
UX & Visual Polish
5/10
Look and feel quality
UX considerations (fallbacks, resilience ladders, honest provenance, fail-safe cron behavior, composited motion) are well articulated in the human-authored architecture, showing a product-minded approach; however, the human-authored artifact is documentation rather than interactive UI code with skeletons or undo patterns.
Evidence
askforai-architecture/index.html: 'Every dependency degrades to a working local behaviour' (resilience ladders)
askforai-architecture/index.html: 'Honest provenance in the UI' and 'Fails safe when unconfigured'
Expertise
Modern Web Frameworks· Junior
React· Junior
Frontend Architecture & Build Tools· Junior
Web Performance & Optimization· Junior
Industries
Artificial Intelligence· Junior
Data & Analytics· Junior
Technologies
JavaScript· since 2026 · Junior
Tailwind CSS
Next.js
Vercel
React.js· since 2025
Cytoscape.js
Element Plus
Recharts
Recommendations
- Lead architecture and build decisions for Next.js apps that need server-first rendering, prerendering and constrained client JS islands.
- Design and implement resilient, cost-constrained systems that fallback cleanly - scheduled jobs, seed fallbacks and access-control patterns.
- Owner for performance and SEO-sensitive product surfaces, driving LCP/SSR budgets and prerender strategies.
- Collaborate on data-heavy dashboards (rendering and aggregation strategy) where server-side aggregation and deterministic client behavior are required.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Intern QA Engineer
Confidence: Low Generalist
A junior full-stack developer focused on analytics dashboards and build-time data processing with a strength in static precomputation for client-side dashboards. The strongest proven skill is build-time aggregation and client decoding of a columnar dataset as implemented in scripts/aggregate.py and public client decoding in lib/dataset.ts. There is no human-authored QA automation, negative-path testing, or CI test-integration code visible for evaluation.
Test Automation Frameworks
Building automated tests
Not evidenced in public code
Test Coverage & Strategy
What and how to test
Not evidenced in public code
API & Integration Testing
Testing how parts work together
Not evidenced in public code
Performance & Load Testing
Testing speed under load
Not evidenced in public code
Bug Reporting & Analysis
Finding and describing bugs
Not evidenced in public code
CI Test Integration
Running tests automatically
Not evidenced in public code
Industries
Lifestyle· Junior
Transportation & Logistics· Junior
Technologies
Python· since 2024 · Junior
TypeScript· since 2025 · Junior
Recommendations
- Implement a small, human-authored test suite around the data pipeline (unit tests for scripts/aggregate.py and src/transform.py) using pytest to lock in correctness.
- Add contract/schema validation for dataset artefacts (JSON meta and binary layout) and a verify test (scripts/verify.py style) into CI so aggregates are reproducible on every change.
- Introduce lightweight QA infra: per-test fixtures, deterministic data factories, and a small set of negative-path tests for anomaly detectors (explicit expected error cases and status checks).
- Integrate verification into CI (single job) with artifact snapshots and a fail-fast policy to protect the precomputed aggregates.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: