MOHD DANISH
Contact details — on requestProof of Work — on request
Experience
P
Junior AI Developer – R&D
Pixis.AI · Mar 2026 – Present
- Built and optimized text-to-image, upscaling, and inpainting pipelines for product ad generation
- — increasing efficiency by 43% and accelerating processing by 30% via component-level queuing and parallel execution.
- Engineered a natural language-driven ad refactoring system, reducing token usage by 37% and decreasing ad discard rates by 23% through single-pass complete ad generation.
- Created an LLM-powered ad generation system using Chain-of-Thought prompting that synthesizes 10 ready-to-use ads within 7 seconds by analyzing web content and existing creatives.
P
Junior AI Developer – R&D
Pixis.AI · Feb 2026 – Present
- Architected a multimodal intelligence system for Facebook/Meta Ads that extracts visual layout, copy structure, CTA patterns, and offer framing from brand creatives to drive data-informed ad generation.
- Developed a production-ready shadow generation pipeline using diffusion and image-to-image techniques, fine-tuned on 25K images built with 3D Unreal Engine - achieving 80% human feedback accuracy for e-commerce ad creatives.
- Designed an SDXL finetuning pipeline with LoRA adapters, deploying 5+ models in production and reducing generation time by 60%.
- Enhanced AI content generation by fine-tuning SDXL, SD1.5, SD2.0, and FLUX models with ControlNet integration, enabling advanced customization and isometric product generation workflows.
- Built MCP servers and agentic pipelines for automated ad generation, real-time campaign insights, and optimization.
- Developed a precise object positioning model for background integration, improving layout quality and workflow automation.
Lead|Programmatic Ad operations and New Initiatives
Times Internet · Sep 2024 – Present
- Team Lead | Display Programatic Ad operation| Video| Revenue Optimization
H
Data Scientist (Part-Time)
Hottrix Pvt Ltd · May 2025 – Dec 2025
- Implemented video proctoring using pose estimation (MediaPipe), real-time cellphone detection with YOLOv7, and face emotion recognition with RepVGG on AffectNet.
- Built fluency assessment tools leveraging Praat and Prosody for pause/articulation analysis, with stacked CNN+LSTM models on spectrogram features for automated speech evaluation.
Assistant Manager
Times Internet · Apr 2022 – Sep 2024
Programmatic Ad Operation
Times Internet · Aug 2020 – Apr 2022
Ad Operations Specialist
NDTV Convergence Limited · Sep 2017 – Aug 2020
Account Executive
Zirca Digital Solutions Pvt Ltd · Sep 2016 – Sep 2017
Skills 0 proven through work
Also works with
Data AugmentationLLM Output Verification and ValidationNatural Language GenerationSemantic Similarity EvaluationPrompt EngineeringQuery WritingDatabase ManagementEmbedding GenerationDocument ChunkingPDF Library UsageLLM Application IntegrationKnowledge Base Design & ManagementData ExtractionChatbot DevelopmentRetrieval-Augmented GenerationLoRA Fine-TuningModel Performance ImprovementConvolutional Neural Network OperationsModel Building from ScratchTechnical R&D and Spike InvestigationResNet Model UsageModel Fine-TuningMultimodal Data ProcessingFreelancingDiffusion Model UsageModel Training ExecutionAI Agent DevelopmentComputer Vision ModelingData ScienceArtificial Intelligence FundamentalsDeep LearningMachine Learning Application
Proof of Work
Proof of Work
MOHD shares this with people who ask. You'll hear back either way.
Education
B.Tech, Computer Science
Maulana Azad National Urdu University · 2018 — 2022
Contact details
Contact details
MOHD shares this with people who ask. You'll hear back either way.
Ask about this candidate
AI responses may contain errors. Proof status and source information are provided by Proof of Skill.
A Proof CV — one profile, kept current, with the proof attached. Back to the portal