Tejaswini Adapa
Fullstack
Management: Under a year (1-5 people)
Node JS
Python
JavaScript
Model Architecture & Training: 4/10
Active 3 days ago
Invite to interview
Download CVCV
Overview
Technical skills
Timeline
Roles
Overview
ML engineer focusing on audio TTS and multimodal avatar pipelines at a middle level with practical implementation strength in model components and integration. The strongest proven skill is implementing and integrating speech synthesis and generative audio modules demonstrated by the SynthesizerTrn and related modules in backend/flask1/ttsmodel/openvoice (models.py and modules.py). There is limited evidence of full training pipelines, experiment tracking, production MLOps, or novel research contributions in the public code.
Phone
Technical skills
Languages
3
Node JS
Python
JavaScript
Node JS
3
Express
Bcrypt
Multer
Frontend
8
React.js
Vite
Next.js
Tailwind CSS
Zustand
React Router
Recharts
i18next
AI/ML
10
NumPy
OpenCV
LLM
Gradio
OpenAI
Pandas
PyTorch
TensorFlow
Transformers
Torchvision
Other
12
PostgreSQL
Flask
MySQL
GitHub
ESLint
Docker
Git
Rest API
Postman
Model Context Protocol
RAG
Prompt Engineering
Timeline
Full Stack Intern
•
Junior
Infosys Springboard
•
Full-Time
Built an end-to-end precision agriculture management platform. Developed the React frontend with login/registration and farmer/crop management screens. Implemented Express.js REST APIs for user authentication and farm operations. Designed PostgreSQL schema for crop and farm records and integrated it with the API layer.
React.js
Vite
Node JS
Express
PostgreSQL
Web Development Intern
•
Junior
Zaalima Development Pvt. Ltd.
•
Full-Time
Developed a resume-building product (CareerForge Pro) using a React/Vite frontend with live split-screen preview. Created a job description analysis component to extract and categorize keywords such as languages, frameworks, tools, and soft skills. Implemented an ATS match score calculator with found/missing breakdowns. Designed a one-click bullet improvement suggestion workflow with rewrite preview.
React.js
Vite
Middle AI/ML Engineer
Confidence: Medium ML Engineer
ML engineer focusing on audio TTS and multimodal avatar pipelines at a middle level with practical implementation strength in model components and integration. The strongest proven skill is implementing and integrating speech synthesis and generative audio modules demonstrated by the SynthesizerTrn and related modules in backend/flask1/ttsmodel/openvoice (models.py and modules.py). There is limited evidence of full training pipelines, experiment tracking, production MLOps, or novel research contributions in the public code.
Model Architecture & Training
4/10
How well models are designed and trained
Custom model components and architectures are implemented (flow modules, attention, TTS generator), but there is little evidence of full training pipelines, hyperparameter management or systematic experiments.
Data Pipeline & Feature Engineering
3/10
How data is prepared for models
Practical data handling and preprocessing for audio, text and images is present, including text cleaning, audio splitting and face encoding, but there is no large-scale data pipeline or feature store evidence.
Evidence
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/text/cleaners.py: cjke_cleaners2 and language-specific text processing
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/se_extractor.py: split_audio_whisper / split_audio_vad
Adapa-tejaswini/face_recognition_attendance/encode_faces.py: dataset traversal, image resizing, face detection and saving encodings.pickle
Experimentation & Evaluation
2/10
How results are measured and tested
Very light experimentation and evaluation artifacts exist (console verification, saved encodings, simple summaries) but no experiment tracking, baselines, reproducible eval scripts or ablation studies.
MLOps & Deployment
2/10
How models are shipped to production
There is basic serving and integration work using Flask and Gradio endpoints and upload routes, indicating initial deployment glue but lacking robust MLOps features like versioning, monitoring, or CI-driven model lifecycle management.
Evidence
Computational Efficiency
3/10
How efficiently computing resources are used
Some computational-efficiency conscious patterns are used, for example depthwise/separable convs, weight normalization and a torch.jit scripted fused kernel, but there is no profiling, quantization results, or distributed/GPU-scaling evidence.
Evidence
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/modules.py: DDSConv uses depthwise separable convolutions
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/attentions.py: @torch.jit.script fused_add_tanh_sigmoid_multiply
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/modules.py: consistent use of weight_norm and remove_weight_norm patterns
Research Depth & Innovation
2/10
Depth of research and new ideas
The codebase implements non-trivial, research-grade model components from known papers and repos, but there is no clear evidence of novel research contributions, reproduced benchmark results, or structured ablation studies.
Evidence
Adapa-tejaswini/Avatar-lab/backend/flask1/ttsmodel/openvoice/models.py: VITS-like SynthesizerTrn architecture
Adapa-tejaswini/Avatar-lab/backend/flask1/lsmodel/models/wav2lip.py: Wav2Lip model implementation
Adapa-tejaswini/Avatar-lab/backend/flask1/lsmodel/face_detection/detection/sfd/net_s3fd.py: S3FD face detector implementation
Verified artifacts
Expertise
Audio & Speech Processing• Middle
Computer Vision & Image Analysis• Middle
Industries
Media & Entertainment• Middle
Technologies
Python• Middle
MySQL
PostgreSQL• since 2025
OpenCV
Model Context Protocol
Prompt Engineering
Gradio
Transformers
TensorFlow
Pandas
NumPy
Git
PyTorch
Docker
LLM
RAG
Torchvision
OpenAI
GitHub
Recommendations
- Productize and document training and evaluation pipelines with reproducible scripts and experiment tracking (WandB/MLflow) for the TTS models.
- Add unit and integration tests plus CI to validate model code, pre/post processing and Flask endpoints to improve reliability.
- Implement lightweight MLOps features: model versioning, simple monitoring/logging and a deployment manifest or Dockerfiles for inference services.
- Reduce maintenance risk by auditing copied/third-party model modules for licensing and adding attribution or refactorings to isolate external code.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Frontend Developer
Confidence: Medium Fullstack
Fullstack web developer at a mid level focused on building agriculture management and related web applications with integrated backend logic and simple frontend state stores. The strongest proven skill is implementing domain-rich server-side business logic and safe data flows, exemplified by controllers for notifications, AI chatbot integration, weather fetching, fertilizer recommendations, and the localStorage-backed client API. Public code shows limited evidence of automated test coverage, CI pipelines, and advanced frontend component system design or measured performance tuning.
UI Component Architecture
2/10
How interface parts are built
Limited evidence of deliberate component boundaries or a custom design system; state management and types are present but UI component architecture is mostly minimal or template-like.
Evidence
task_manager/store/useTaskStore.ts: Zustand persistent store used to manage task state
task_manager/types/task.ts: explicit Task and TaskState TypeScript types
harry-potter-movie-explorer/script.js: modular DOM functions for rendering and modal handling (no React components)
Responsive & Cross-browser
4/10
Works on all screens and browsers
Responsive CSS and basic cross-browser considerations are present via media queries and viewport meta plus prefers-color-scheme support, but no advanced layout strategies or RTL/i18n plumbing were found in UI code shown.
Evidence
harry-potter-movie-explorer/styles.css: responsive grid, media queries and adaptive modal layout
harry-potter-movie-explorer/index.html: meta viewport and semantic sectioning
task_manager/app/globals.css: prefers-color-scheme dark mode variables
Performance Optimization
2/10
Speed of the interface
Little to no evidence of measured performance work, bundle analysis, code splitting, virtualization or targeted render optimizations; basic caching via localStorage exists but no profiling artifacts or advanced optimizations.
Evidence
farmverse_agriculture_management/src/api.js: localStorage-backed reads and writes used as a simple offline/cache layer
server/controllers/chatbotController.js: model fallback loop to handle busy or rate-limited AI models (operational resilience rather than measured performance tuning)
Accessibility & Semantics
3/10
Usable for everyone
Some accessibility and semantics awareness in static pages and keyboard handling is present, but accessibility on custom widgets and focus management in complex UIs are limited.
Evidence
harry-potter-movie-explorer/index.html: semantic tags and meta attributes
harry-potter-movie-explorer/script.js: keyboard support for Escape to close modal and overlay click handling
State Management & Data Flow
5/10
Managing data in the app
Clear server- and client-side state discipline with validation, idempotent upserts, server-side record validation, token-based auth headers, and a persistent client store; demonstrates considered data flow and backend business logic handling though no formal state machines or request cancellation patterns were present.
Evidence
farmverse_agriculture_management/src/api.js: local token handling, authHeaders, validations, duplicate checks, and collective order allocation logic
task_manager/store/useTaskStore.ts: persistent zustand store with actions and typed state
server/controllers/chatbotController.js: getValidRecordRefs, generateWithFallback and structured DB inserts for chatbot history and scans
UX & Visual Polish
4/10
Look and feel quality
Good visual polish on static UI (animations, transitions, responsive grid and hero sections) and pragmatic UX choices, but UX consistency across a React frontend and comprehensive edge state UI patterns are not fully evidenced.
Evidence
harry-potter-movie-explorer/styles.css: animations, hero section, card hover states and responsive adjustments
task_manager/app/globals.css: theme variables and readable base styles
Expertise
React• Middle
Modern Web Frameworks• Middle
Frontend AI Integration• Junior
Industries
Farming & Agriculture• Middle
Technologies
JavaScript
Node JS• since 2025 • Middle
Zustand
Tailwind CSS
Next.js
Express• since 2025
Bcrypt
Multer
React.js• since 2024
Vite• since 2024
ESLint
Recharts
i18next
React Router
Recommendations
- Build small to mid-size agritech features that need domain logic, backend APIs, and data validation such as notifications, scheduling, and recommendations.
- Implement end-to-end features that connect a React/Vite frontend to Express/Postgres APIs, especially where AI-assisted workflows are required.
- Own feature work that combines server-side integration (external APIs, model fallbacks) with pragmatic client-side persistence and state management.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Middle Game Developer
Confidence: High Generalist
Full-stack web developer (middle level) focused on building practical agriculture management services with backend APIs and frontend state tooling. The strongest proven skill is backend application design and domain logic for agritech workflows, demonstrated by controllers implementing smart notifications, fertilizer recommendations, weather integration, and AI-backed chatbot/image analysis (server/controllers/*.js). The public code does not show any game development engine, rendering, or deterministic simulation work and lacks automated performance profiling or unit test coverage.
Gameplay Systems & Mechanics
How game logic works
Not evidenced in public code
Graphics & Rendering
Drawing game visuals
Not evidenced in public code
Physics & Math
Game physics and math
Not evidenced in public code
Engine Proficiency
Skill with the game engine
Not evidenced in public code
Performance & Frame Budget
Keeping the game smooth
Not evidenced in public code
Content Pipeline & Tooling
Tools for game content
Not evidenced in public code
Industries
Farming & Agriculture• Middle
Recommendations
- Develop full-stack agritech features - backend APIs, PostgreSQL data models, and frontend integrations for farm and crop management.
- Implement AI-assisted features that need safe guardrails - image analysis pipelines, chat assistants, and model fallback strategies.
- Build and maintain domain-specific utility services - notification engines, scheduling/advisory logic, and collective marketplace allocation.
- Harden production readiness - add automated tests, CI, and explicit error/edge-case handling around external API failures.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: