Creative Producer & Creative Technologist | Generative AI, Commercials, Branded Content & Visual Production

Andriy Kostyuk • Peerlist

The Rising City is a cinematic generative AI science-fiction sequence built around the emergence of an ancient alien metropolis from beneath the ocean.The film begins at water level as a single monumental central tower breaks through the surface, displacing huge volumes of water and revealing the first glimpse of a much larger structure hidden below.From there, the sequence expands underwater and above the surface, gradually revealing a connected radial city built around six massive districts, bridges, terraces and hundreds of smaller towers. Rather than having structures grow or materialize independently, the entire metropolis rises as one coherent architectural system.A major focus of the concept was scale and physical continuity — maintaining a consistent city layout while dealing with water displacement, cascading waterfalls, underwater currents, foam, mist, atmospheric perspective and the changing relationship between the architecture and the ocean.The visual language combines dark silver and bronze materials with restrained cyan energy lines, creating a sense of ancient technology gradually coming back to life.The sequence builds toward a final aerial reveal, where the complete city stands above the ocean as light spreads outward from the central tower across the six districts.The project explores generative AI for large-scale cinematic world-building, coherent environment design and VFX-driven science-fiction imagery.

Atlantis: Rise from the Deep

Bring the Noise is a high-energy generative AI dance film combining grounded hip-hop choreography with bold dimensional motion design.The piece follows two consistent dancers across saturated violet, acidic lime, hot tangerine and black studio environments, moving through fast footwork, deep bounce, heel-toe switches, running-man variations, chest hits and synchronized combinations.The visual system is built around oversized 3D typography and sculptural graphic elements that respond directly to the choreography. Words such as “BOUNCE,” “HEAVY,” “HIT,” “LOUD” and “BRING THE NOISE” become physical parts of the environment rather than flat overlays, landing, stretching and reacting to movement on the beat.A major focus of the project was keeping the dancers’ motion readable and physically grounded while allowing the graphics to remain equally energetic. Camera moves stay short and controlled, feet remain visible during choreography, and the dimensional forms are designed to support rather than obscure the performance.The film moves between solo sequences, graphic inserts and synchronized duet sections, building toward the final “BRING THE NOISE” composition while keeping the energy live all the way through the final frame.The project explores how generative AI can combine choreography, character consistency, 3D typography and music-driven motion design within a compact music-video language.

Bring the Noise

The Ashen Atlas is a cinematic generative AI title sequence built around the idea of an ancient mechanical map slowly awakening into a living fantasy world.The sequence begins with a monumental circular atlas carved into black stone, where engraved coastlines, copper borders and glowing waterways gradually come to life.As the camera moves closer, the flat cartography transforms into a physical miniature landscape. Cliffs rise from layered terrain plates, bridges extend across fissures, towers assemble from mechanical structures, and an observatory unfolds through gears, rails and rotating brass components.The visual language combines dark fantasy, cartographic design and intricate mechanical detail with the tactile feeling of a handcrafted practical miniature. Stone, oxidized copper, bone, brass and molten-gold light create a restrained material palette while the slow camera movement reinforces the sense of scale.Rather than revealing the title immediately, the environment itself gradually constructs its identity. Rivers, terrain and mechanical elements converge into a central emblem before the illuminated cartographic lines detach from the atlas and assemble into the final “THE ASHEN ATLAS” title.The project explores generative AI as a tool for cinematic title design, fantasy world-building and highly detailed environment-driven visual storytelling.

The Ashen Atlas

Built for Impact is a generative AI American-football campaign combining athletic performance, fashion-editorial styling and aggressive motion design.The film follows three consistent athletes through a tightly controlled black-and-crimson visual world, moving between preparation portraits, sprinting, catches, route running, blocking and full-contact play.The graphic language is treated as part of the action rather than as a layer placed on top of it. Oversized typography, barcodes, route lines, field coordinates, counters, crosshairs and technical markings react to movement, map plays and carry momentum between shots.A key part of the concept was balancing two visual identities at once: the physical weight and intensity of American football, and the precision of a premium fashion campaign. The athletes remain visually consistent while the film moves rapidly between beauty portraits, detailed equipment shots, high-speed action and impact-heavy sports cinematography.The sequence builds around themes of power, speed, control and attitude, culminating in the final “BUILT FOR IMPACT” hero composition.The project explores how generative AI can combine character consistency, complex athletic action, editorial art direction and dense motion graphics within a short-form commercial format.

Built for Impact

LUMA - Charge the Moment is a stylized generative AI beverage commercial built entirely as a 2D motion-graphics piece.&nbsp;Unlike a photoreal product ad, this version explores a flatter, more graphic visual language built from bold outlines, geometric shapes, high-contrast color, halftone accents and fast rhythmic animation.&nbsp;The film is centered around a fictional LUMA beverage can and develops its identity through electric yellow, neon lime, deep violet and crisp white, using motion design as the main storytelling tool. Concentric rings, lightning bolts, checkerboard bands, arrows, bubbles and energy waves all work together to create a playful, high-impact campaign world.&nbsp;A key focus of the project was translating the feeling of carbonation, speed and refreshment into a purely graphic language without relying on photoreal liquid simulation or traditional product cinematography.&nbsp;The spot moves from an energetic graphic reveal into can rotation, pull-tab action, abstract motion sequences and a final packshot built around the slogan “CHARGE THE MOMENT.”&nbsp;The result is a short-form commercial experiment exploring how generative AI can be used not only for cinematic or photoreal visuals, but also for stylized motion-design-driven brand storytelling.Tool used: MiniMax H3

LUMA - Charge the Moment

LUMA - Citrus Fizz is a generative AI beverage commercial built around bold color, summer energy and graphic-driven product storytelling.The piece combines high-speed product cinematography, pop-fashion styling and Memphis-inspired motion graphics to create a playful, premium commercial language around a fictional citrus soda brand.Visually, the film moves between macro product detail, poolside fashion performance, stylized liquid worlds and fast split-screen editorial layouts, all tied together by a bright palette of lemon yellow, cobalt blue, cyan and hot pink.The central challenge was balancing the commercial clarity of the product with a more expressive visual identity — allowing the can, its condensation, carbonation, citrus elements and liquid motion to remain the hero while the surrounding graphics added rhythm, attitude and memorability.Typography, geometric shapes, halftone textures, checkerboard accents and animated wave forms were treated as part of the campaign world rather than simple overlays, helping the spot move between product beauty shots and more energetic lifestyle moments.The result is a short-form generative AI commercial exploring how product visualization, graphic design, fashion styling and motion can come together in a bright, contemporary advertising format.

LUMA - Citrus Fizz

Own Your Energy is a high-energy generative AI fashion-sport film combining streetwear, athletic movement and bold editorial motion design.The piece follows one consistent character through a sequence of fast, graphic-driven environments built around an electric-lime, coral, off-white and charcoal palette.Rather than treating typography as a simple overlay, the graphic system becomes part of the action itself. Oversized words, cropped letterforms, arrows, grids, halftone textures, poster fragments and directional elements react to the character’s movement and reshape the frame from shot to shot.The performance combines sprint starts, pivots, jumps, slides and short athletic gestures with fashion-editorial posing, keeping the movement energetic without losing the sense of a premium campaign.A major part of the concept was maintaining a consistent character, wardrobe and visual identity while shifting rapidly between very different compositions and motion-graphic treatments.The film builds toward the final “OWN YOUR ENERGY” composition - a clean editorial hero frame that brings the movement, typography and fashion direction into one visual statement.The project explores how generative AI can bring together character consistency, athletic motion, fashion styling and kinetic graphic design within a short-form commercial language.

Own Your Energy

Own the Frame is a generative AI fashion film combining editorial styling, bold typography, geometric set design and controlled motion. The piece follows one consistent character through a sequence of graphic fashion environments built around a restrained palette of deep blue, orange, cream, black and metallic accents. Typography becomes an active part of the visual direction, with statements such as “MOVE DIFFERENT,” “NO LABELS,” “OWN THE FRAME” and “WEAR THE MOMENT” integrated directly into the compositions rather than added as simple overlays. The visual language moves between clean studio performance, oversized architectural shapes, chrome elements, portrait collage and graphic transitions, while maintaining a consistent character identity and fashion direction across the sequence. The focus was on balancing model performance, styling, camera movement and motion graphics without letting any one element overpower the frame. The result is a short-form fashion piece exploring how generative AI can combine character consistency, art direction, typography and motion design within a polished editorial language.

Own the Frame

Style in Motion is a high-energy generative AI fashion film built around bold color, editorial styling, kinetic typography and fast graphic transitions.The piece follows one consistent character through a series of distinct fashion looks and visual worlds - from sporty cobalt and cyan to streetwear, lavender editorial sets and a final clean hero look.Rather than treating typography and graphics as decoration, the visual system becomes part of the performance. Oversized words, magazine layouts, ribbons, portrait cards, geometric shapes and color fields constantly reorganize the frame around the character and her movement.A major part of the concept was maintaining the same identity across different outfits, hairstyles, environments and graphic treatments while keeping each section visually distinct.The pacing moves between fast editorial cuts, controlled camera movement, graphic wipes and one deliberate quiet break before returning to the final “STYLE IN MOTION” composition.The result is a short-form fashion piece exploring how generative AI can combine character consistency, styling, motion design and graphic direction within a single visual language.

Style in Motion

Prehistoric Games is a generative AI sports reel built around an unlikely collision: modern action sports set against a prehistoric world.The project moves through a series of independent scenarios - motocross riders racing alongside dinosaurs, kayaking through wild rivers, hang gliding with prehistoric flyers, mountain biking through rugged terrain, snowboarding across alpine landscapes, tennis, track and field, and stadium action surrounded by creatures from another era.The focus was on combining fast human movement with large-scale creatures and environments while keeping each scene readable, energetic and visually coherent. Different sports introduced different challenges: speed, camera tracking, interaction with terrain, water, dust, snow, scale and the physical presence of creatures moving through the same space.Rather than building a single narrative, the reel explores a range of visual situations connected by the same central idea - familiar sports transformed by the presence of a prehistoric world.Tools used: ChatGPT Images 2.0 - image generation LTX 2.3 - image-to-video generation ACE STEP 1.5-XL - music generationPart of an ongoing exploration of generative AI for action, sports and cinematic visual production.

Prehistoric Games

Beyond Earth is a generative AI space reel exploring scale, motion and cinematic atmosphere across orbital and lunar environments.The piece moves through a series of spaceflight scenarios - orbital stations, astronaut EVA, rocket launches, docking sequences, spacecraft operations and lunar surface exploration.The focus was on creating a convincing sense of scale and physical presence across very different shots: large orbital structures, mechanical detail, camera movement, propulsion, weightlessness and the contrast between Earth orbit and the lunar landscape.Rather than building around a single hero shot, the project explores how generative AI can maintain a coherent cinematic language across a wider range of environments, vehicles and types of motion.Tools used: LTX 2.3 — video generation ChatGPT Image 2.0 — image generation ACE Step 1.5 XL — music generationPart of an ongoing exploration of generative AI for cinematic visual production, world-building and large-scale action sequences.

Beyond Earth

A narrative music video created with a fully local generative AI production workflow.The project was built around a simple challenge: creating not just a series of visually strong shots, but a coherent music video with consistent characters, visual language, environments, styling and camera direction across the entire sequence.Character consistency was supported with custom LoRAs for both image generation and LTX 2.3, while the wider production challenge was maintaining continuity across framing, movement, lighting, wardrobe, performance and mood.The video was produced locally on a single workstation, using:LTX 2.3 -&nbsp;video generation Z Image Turbo, FLUX.2 [klein], Qwen Image Edit -&nbsp;image generation and editing ACE-Step 1.5 -&nbsp;music generation ChatGPT 5.4 -&nbsp;lyrics and prompt developmentWhat interested me most in this project was the shift from generating individual shots to directing a complete sequence. A strong standalone image or clip is relatively easy to judge in isolation; maintaining a consistent creative language across an entire piece is a very different production problem.The project is part of my ongoing exploration of how generative AI can be integrated into real filmmaking workflows -&nbsp;not only for rapid visual experiments, but for complete narrative and commercial production.

Breaking Through (Music Video)

A generative AI vehicle action study exploring speed, motion and cinematic energy across a range of driving scenarios.This project brings together a series of high-energy automotive moments -&nbsp;drifting cars, rally action, off-road movement, racing shots and monster trucks -&nbsp;with the aim of exploring how generative AI can handle vehicles as fast, dynamic subjects.Rather than focusing on a single hero shot, the project looks at a wider range of motion challenges: different types of vehicles, different environments, camera movement, dust, weight, traction and the interaction between the vehicle and the surrounding space.The main creative interest here was not only the cars themselves, but the sense of momentum around them -&nbsp;how the camera follows the action, how the environment reacts, and how much physical energy and cinematic intensity can be built through generated imagery and motion.Tools used: LTX 2.3 -&nbsp;video generation ChatGPT Image 2.0 -&nbsp;image generation ACE Step 1.5 XL -&nbsp;music generationThis project is part of an ongoing exploration of generative AI for automotive and action-focused visual work, with a particular interest in movement, rhythm, speed and screen presence.

Generative AI Automotive Reel

Make It Yours is a 15-second generative AI fashion concept built around the collision of streetwear, live-action movement and oversized editorial graphics.The piece follows one character through three distinct looks, moving between direct-flash fashion imagery, photocopied cutouts, torn-paper transitions, kinetic typography and full-screen magazine-style layouts.Rather than treating graphics as decoration, the visual system constantly reshapes the frame - freezing, replacing, pushing and tearing through the live-action image in sync with the character’s movement.The campaign uses a tight palette of cherry red, butter yellow, dirty white, ink black and brushed silver, with a tactile editorial finish inspired by photocopies, printed matter and fashion publishing.Fast cuts, physical gestures and graphic transitions are synchronized to a 145 BPM club track, building toward the final “MAKE IT YOURS” statement.A short-form experiment in how generative AI can be used not only to create imagery, but to build a complete visual language across fashion, motion and graphic design.

Make It Yours

Invisible People is a surreal generative AI video experiment built around one simple visual idea: remove the human body while preserving the sense of its physical presence.The film is made up of a series of independent everyday moments - invisible dancers, a tennis player, street musicians, a cyclist, a woman walking an invisible dog, shoppers and other characters moving through familiar environments. Their bodies are absent, while clothing, accessories and movement still make their presence immediately readable.The project was created through an image-to-video workflow. Still images were generated with GPT Image 2.5 and then animated with MiniMax H3.One of the main challenges was maintaining a consistent visual rule across very different scenes: clothing needed to retain believable volume and interior surfaces, while avoiding transparent skin, glass-like human forms, refraction and other artifacts that would break the illusion.The result is a playful, slightly surreal exploration of generative AI video, invisible characters and visual storytelling - turning ordinary situations into scenes where the people are missing, but their presence remains.